CN102867414A - Vehicle queue length measurement method based on PTZ (Pan/Tilt/Zoom) camera fast calibration - Google Patents

Abstract

The invention discloses a method for measuring vehicle queuing length based on rapid calibration of PTZ cameras. , to establish the conversion relationship between the coordinates of the pixels in the image and the coordinates of the corresponding points on the road surface in the world coordinate system; use PTZ cameras to obtain video images of traffic monitoring scenes, set the ROI of the lane, and use the background adaptive update algorithm and texture The feature detects the vehicle queuing status in the domain of interest, and obtains the pixels and coordinates of the tail of the queuing vehicles; converts the detected pixel coordinates of the tail of the queuing into world coordinates, and finally calculates the length of the vehicle queuing. The invention judges the tail position of the vehicle queuing according to the texture feature, and completes the measurement of the vehicle queuing length in combination with the camera calibration, and has the advantages of low cost, strong embeddedness and the like.

Description

A Method for Measuring Vehicle Queue Length Based on Rapid Calibration of PTZ Camera

technical field

The invention relates to a vehicle queuing length measurement method, in particular to a vehicle queuing length measurement method based on rapid calibration of a PTZ camera.

Background technique

With the rapid development of the economy, more and more cars are owned in cities, and traffic congestion has always been one of the problems in urban development. Intelligent transportation system is one of the effective ways to solve urban traffic congestion. At present, most traffic intersections in China adopt timing control of traffic lights, and the time length of traffic lights switching is fixed. Due to the complexity and randomness of traffic flow changes, the defects of this fixed allocation method are becoming more and more prominent. Therefore, research on the transition from fixed allocation time to intelligent time allocation has been gradually carried out. When the urban traffic signal command system performs intelligent time allocation on the intersection signal, it needs to obtain the traffic flow information of the road. Among them, the queue length of vehicles at the intersection is one of the important parameters in the intelligent transportation system.

The main ways to obtain traffic flow information include ground sensing coils, video recognition, and floating car estimation. Since the ground induction coil is placed at a fixed position on the road, although the number of vehicles can be accurately measured, there is a large error in measuring the length of the vehicle queue. The patent with the publication number CN102024323A discloses a method for extracting vehicle queuing length based on floating car data. The method first extracts the vehicle-mounted GPS of the floating car that is normally queued up before the intersection of the road section through road section matching technology to obtain the stop point position, and then The position distribution changes of the floating car stop point and the intersection are counted to estimate the queue length.

The intelligent traffic monitoring system based on video analysis has the advantages of obtaining more traffic information, a large monitoring range, and easy installation. Researchers at home and abroad tend to use computer vision technology to automatically extract vehicle length. Zhu Xiaoshan and others proposed a vehicle queue length detection method that improves the Canny edge detection algorithm. Rourke and Bell proposed an FFT-based traffic queue detection method. Li Weibin and others used gradient difference and color difference to obtain complete vehicle information, and used telescopic windows to detect queue length. The patent with the publication number CN101936730A discloses a method and device for detecting the length of vehicle queuing. During the day, the three-frame difference method and morphology are used, and at night, the vehicle light characteristics are used to detect the queuing vehicle queue.

Scholars at home and abroad mainly focus on the pixel information of vehicle queuing detection in images through video analysis, but there are few studies on the calculation of vehicle queuing length combined with camera calibration, especially PTZ (Pan/Tilt/Zoom) cameras, because of their frequent internal and external parameters. It is more difficult to calculate the length of the vehicle queuing queue.

Among the existing calibration methods, the traditional camera calibration method usually uses a specific calibration block, and the calibration point setting process is cumbersome. The self-calibration method needs to control the surveillance camera to perform several rigid-body motions and perform feature parameter matching on images under different viewing angles, which is not suitable for applications where surveillance cameras are usually fixed in traffic scenes.

Aiming at the characteristics of the traffic surveillance scene, Nelson, Granttham and George et al. adopted the camera calibration method of the traffic video surveillance system based on the simple imaging model, and directly used the rectangle composed of the corner points of the road lane boundary line in the traffic scene to realize the calibration of the camera focal length and orientation parameters. . Invention CN1564581A discloses a method for calibrating camera focal length and space external parameters by using several special straight lines on the road as calibration targets in the urban traffic monitoring environment. The above-mentioned method is based on the three-dimensional mapping model, and some internal parameters of the camera model need to be solidified, and many known conditions are required, which reduces the scope of use of the calibration algorithm.

Fathy et al. used the method of camera calibration to establish a conversion model from 2D image coordinates to 3D world coordinates to calculate the length of the queue. In order to calibrate the parameters of the camera, this method needs to know the coordinates of four points in the world coordinate system and the coordinates of the corresponding points in the image coordinate system in advance. The acquisition of coordinate points is inconvenient, and the practical application is troublesome, and it is not easy to obtain such coordinate points in actual scenes, and additional calibration reference objects need to be placed. Fung et al. used the existing lane lines to complete the camera calibration, and only needed to know the width of the lane. Although the lane width follows a certain standard and is easy to obtain, the traffic markings in this algorithm must have rectangular features, and there are certain limitations in adaptability. sex. Li Bo and others proposed to use two line segments collinear of road markings and another parallel line as a calibration template, but the height of the camera must be known. The above methods have defects such as difficult selection of reference markers with specific geometric features and difficult access to constraints, which have certain limitations in practical application.

Contents of the invention

In order to solve the above-mentioned technical problems existing in the existing vehicle queuing length measurement method, the present invention provides a vehicle queuing length measurement method based on PTZ camera rapid calibration.

The technical scheme that the present invention solves the problems of the technologies described above comprises the following steps:

1) Obtain the video image of the traffic monitoring scene in real time with the PTZ video camera;

2) Select two vertically intersecting traffic markings in the image to form a T-shaped camera calibration reference object, obtain the pixel coordinates of the four endpoints in the T-shaped calibration reference object, and measure the distance between two of the three collinear points , and measure the distance from another point to the line;

3) Establish the model and corresponding relationship between the image coordinate system and the world coordinate system, solve the internal and external parameters of the camera, and establish the conversion relationship between the coordinates of the image pixel points and the coordinates of the corresponding points on the road in the world coordinate system;

4) In the image, select the vehicle queuing information detection area from the stop line of the lane to the preset lane detection area, that is, the region of interest ROI, and the subsequent processing is limited to the ROI; for any pixel (x, y), The Tamura texture feature of its neighborhood image block is used to measure the roughness of the pixel, which is recorded as F _crs (x, y), and F _crs (x, y) is used as the pixel gray value of the texture background image; Calculating the neighborhood roughness value F _crs (x, y) of each pixel to form a background image B _crs representing texture features, and the background image can be adaptively updated as the sunlight changes;

5) Compare the texture features of the real-time image with the texture features of the background image, extract the foreground image, and obtain a binary foreground image with smooth edges and no holes through morphological processing, that is, vehicle queuing information, from the lane stop line to In the lane queuing direction, use the algorithm of moving the detection window to search for the tail position of the queue, and obtain the image coordinates of the pixel N at the tail of the queue;

6) According to the coordinate system conversion relationship established by camera calibration, the image coordinates of the pixel N at the end of the queue are converted into the coordinate values of the corresponding points on the road surface in the world coordinate system, and the queue length of vehicles queuing up in traffic jams is calculated.

The present invention extracts the static background in the video, adopts the texture feature, separates the non-background object or the moving object in the video, and after judging the queuing state of the vehicles in the lane, obtains the internal and external parameters according to the camera calibration, and establishes the world coordinate system and the image coordinate system The conversion relationship of the corresponding points in the center can finally realize the visual measurement of the vehicle queuing length. The invention has the advantages of easy selection of a camera calibration target, additional placement of calibration reference objects, easy operation of the calibration method, good robustness of vehicle queuing state identification, and the like.

The present invention has the following advantages:

1. In the monitoring scene, there is no need to set additional camera sign reference objects, and the geometric characteristics and size data of existing traffic markings are fully utilized. The calibration method of camera parameters is simple and convenient.

2. The invention adopts the texture feature to extract the vehicle queuing information in the background image, avoids the instantaneous change of the illumination caused by the vehicle passing the camera, and adopts the method of automatically updating the background, which can adapt to the change of the sunshine light in different periods.

3. The invention makes full use of the video image information collected by the existing surveillance cameras at the traffic crossing, and is easy to embed into the traffic information management platform, and has the advantages of low cost and strong embedding.

Description of drawings

Fig. 1 is a flowchart of the present invention.

Fig. 2 is a schematic diagram of taking traffic markings as reference objects for camera calibration in the present invention.

Fig. 3 is a camera model diagram in the present invention.

FIG. 4 is a schematic diagram of the calibration reference object and its positional relationship in the camera model in the present invention.

Fig. 5 is a schematic diagram of auxiliary lines in camera calibration according to the present invention.

Fig. 6 is a schematic diagram of an embodiment of the present invention.

Detailed ways

Measuring principle of the present invention is as follows:

(1) Select two traffic markings that intersect vertically to form a T-shaped calibration reference object, and measure the distance between two known collinear three points, and the vertical distance from a point outside the line to the line to obtain four feature points The pixel coordinates in the image, and then solve the internal and external parameters of the camera calibration, and establish the conversion relationship between the point coordinates in the world coordinate system and the pixel coordinates in the image coordinate system;

(2) Extract the image sequence in the monitoring video of the traffic scene, extract the texture features of the background image in the lane area, and establish the texture background image; set the interest domain of the lane in the image, extract the texture features of the lane interest domain in the real-time image, and The texture features of the corresponding area in the background image are automatically compared and judged, and the position of the tail of the vehicle queue in the image is searched;

(3) According to the pixel position of the tail of the searched vehicle queue, the scalar is converted into the point coordinates of the world coordinate system, and finally the length of the vehicle queue in the world coordinate system is calculated.

The specific steps of the above measurement principle are:

(1) Quick calibration of the camera

The present invention adopts a pinhole camera model, assumes that the traffic road surface in the traffic monitoring scene is smooth and has no curl, and ignores the distortion of the camera. The internal and external parameters of the camera include the focal length f, the tilt angle θ, the rotation angle φ, and the angle α between the lane and the horizontal axis of the world coordinate system.

Step 1: Establish models of the image coordinate system and the world coordinate system.

World coordinate system X _W -Y _W : X _W axis is defined as the intersection line between the plane passing through the camera lens center and parallel to the CCD sensor and the ground, and Y _W axis is defined as the intersection line between the plane passing through the camera optical axis and perpendicular to the ground and the ground The positive direction of the Y _{and W} axes is pointing forward along the road surface, and the positive direction of the X _{and W} axes is pointing to the right horizontally; its origin O _W is defined as the intersection of the X and _W axes, as shown in Figure ₃ .

Image coordinate system XOY: In the image, the size of the image is set to M×N, and the origin is defined as the geometric center of the image [(M-1)/2, (N-1)/2]. In the image, the X-axis points horizontally to the left, and the Y-axis points vertically downward, as shown in Figure 3.

When the camera looks down at the calibration reference object in the traffic scene, the angle between the optical axis and the road plane is θ. The angle between the lane dividing line and the world coordinate system is α. The rotation angle of the camera around its optical axis is φ, and when the camera is installed horizontally, the rotation angle φ=0. When the camera is installed non-horizontally, there is an angle between the bottom edge of the camera CCD and the horizontal line, that is, the rotation angle φ is not 0, and the image needs to be rotated by φ to correct the error caused by the non-horizontal installation of the camera.

Step 2, select a reference calibration object. Since the traffic markings such as the edge line of the roadway and the dividing line of the roadway have specific dimensions and geometric shapes, the three collinear endpoints A, B, and C on the dashed line of the lane dividing line are selected for the traffic markings in the traffic scene , assuming that the coordinates in the world coordinate system X _W Y _W are (X _A , Y _A ), (X _B , Y _B ) and (X _C , Y _C ), respectively. Measure the distance between two endpoints, the distance between endpoints AB is h ₁ , and the distance between endpoints AC is h ₂ . Take a certain point D on the vertical line of the dotted lane dividing line through a certain end point, and measure the distance d ₁ between this point and the intersection point. According to the standard specification, the values of h ₁ , h ₂ and d ₁ can be found from the manual or actually measured.

Obtain the pixel points corresponding to the three endpoints A, B, C and vertical point D of the lane dividing line in the image, which are respectively a, b, c and d, and the coordinates in the image coordinate system XY are (x' _a , y ' _a ), (x' _b , y' _b ), (x' _c , y' _c ), and (x' _d , y' _d ).

Step 3: Make an auxiliary line in the image to obtain the data required for camera calibration.

When the camera is not installed horizontally, there is a rotation angle φ between the bottom edge of the CCD sensor and the horizontal line. The value of the angle φ can be calculated by the formula described later. In view of the influence of the rotation angle φ, in order to meet the method described in the present invention, it is necessary to rotate the acquired image by the angle φ correspondingly before calibrating the camera, so as to correct the error caused by the non-horizontal installation of the camera, that is, the rotation of a, b, c and d After the coordinate value is

$\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}\\ \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}\end{array}\right]\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\\ {\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\end{array}\right]<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Calculate intermediate variables $Y_A^{\&prime;}=\frac{h_1{h}_2({\mathrm{the\; y}}_b-{\mathrm{the\; y}}_c)}{({\mathrm{the\; y}}_a-{\mathrm{the\; y}}_b)h_2-({\mathrm{the\; y}}_a-{\mathrm{the\; y}}_c)h_1}---\left(2\right)$

$\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

In the world coordinate system, the point C of the crossing lane dividing line is drawn parallel to the X _w axis, and the point A of crossing the lane dividing line is drawn parallel to the Y _w axis, and the two intersect at point A'. The point A' is projected into the image is a', and its coordinates are (x _a ', y _c ), where

${\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

In the image, a horizontal line is drawn through point d, and the point of intersection with the straight line where a, b, and c are located is point e. The pixel coordinates of point e are (x _e , y _e ), as shown in FIG. 5 .

Step 4: Calculate the angle α between the lane boundary line and the horizontal axis of the world coordinate system.

According to the coordinate values of the four points a, b, c, and d obtained by the image rotation angle φ, the angle α between the lane boundary line and the horizontal axis of the world coordinate system satisfies the following expression.

$\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; e</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; |</span>}$

In the above formula x _e is the abscissa of point e in the image plane, x _a is the abscissa of point A, h ₂ is the distance between two points AC, and d ₁ is the distance between two parallel lines on the road surface.

but

Therefore according to $\sin \mathrm{\&alpha;}=\sqrt{\frac{1-\cos 2\mathrm{\&alpha;}}{2}},$ have to $\mathrm{\&alpha;}=\arcsin \left(\sqrt{\frac{1-\cos 2\mathrm{\&alpha;}}{2}}\right)---\left(5\right)$

Step five, solve the rotation angle φ of the camera extrinsic parameters.

After calculating the angle α between the lane boundary line and the horizontal axis of the world coordinate system, convert the corresponding pixel coordinates of the calibration reference point D in the image into the coordinates of the world coordinate system. The calculation formula is as follows.

${\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}<span\; class="notranslate">\; \&Center\; Dot;</span>\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

Assuming that the straight line passing through point D and the straight line L ₁ intersect perpendicularly at point B, the coordinates of pixel b are converted into coordinates of the world coordinate system in the same way. Finally, the distance of |DB| is calculated, which is the distance from point D to the straight line L ₁ where the three points A, B, and C are located.

In order to solve φ, rotate the image according to formula (1) in the interval φ∈[-15°, 15°], and repeat steps 3, 4, and 5. Use the Levenberg-Marquardt algorithm to search in the interval φ∈[-15°,15°], so that

φ=argmin(|DB|) (8)

Step 6, solve the angle of view θ of the external parameters of the camera

After solving the rotation angle φ, obtain the coordinate values (x _a , y _a ), (x _b , y _b ), (x _c , y _c ) and (x _d , y _d ).

calculate $\sin \mathrm{\&theta;}=\mathrm{\&lambda;}\frac{d_1}{\sin \mathrm{\&alpha;}}\frac{1}{x_e-x_a}\frac{1}{Y_{A^{\&prime;}}\sin \mathrm{\&alpha;}},$ Therefore, the viewing angle θ of the camera can be solved according to the following formula.

$\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; arcsin</span>}<span\; class="notranslate">\; [</span>\frac{{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}{{\mathrm{<span\; class="notranslate">\; sin</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; e</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; Y</span>}}_{{\mathrm{<span\; class="notranslate">\; A</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

Step seven, solve the focal length f of the camera

According to the obtained overlooking angle θ and the value of the intermediate variable λ, the focal length f of the camera can be calculated.

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}{\mathrm{<span\; class="notranslate">\; the\; tan</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

(2) Video image analysis of vehicle queuing status

In order to extract the lane vehicle queuing situation in the video, the present invention uses texture features to extract moving objects from the background image, and the specific steps are as follows:

Step 1: Set the vehicle queuing detection area in the image, that is, the domain of interest ROI, and select the detection area from the traffic stop line in the lane to the predetermined lane detection area as the vehicle queuing detection area. For each frame image of the video, only the image data is calculated for the set detection area.

Step 2, calculate the roughness of each pixel neighborhood in the image. Assuming that the neighborhood of pixel G(x,y) in the image is a window of 2 ^k × 2 ^k pixels, the roughness of the neighborhood image block is calculated using the Tamura texture feature method, denoted as F _crs (x, y), and Calculate the roughness of the neighborhood image blocks of each pixel in the detection area one by one, and form a new image with the neighborhood roughness value F _crs (x, y) of each pixel as the texture feature background image B _crs .

Step 3, using Gaussian model analysis to extract the foreground image F _img from the background image.

In the above announcement, B _crs (x, y) represents the roughness value of the pixel (x, y) in the texture background image, and F _crs (x, y) represents the roughness of the neighborhood block where the current pixel (x, y) is located . If the Gaussian probability of the pixel is greater than the threshold T, it means that the pixel is a foreground pixel.

If there are some isolated points and local holes, the morphological processing method is used to perform corrosion processing and then dilation processing to obtain a smooth and complete foreground image, that is, a binary image of the vehicle queuing.

Step 4: From the lane stop line to the lane queuing direction, use the algorithm of the moving detection window to search for the tail position of the queue, and obtain the pixel coordinates N(x' _n , y' _n ) of the queue tail.

Step 5: Take the current frame _f1 and the previous two frames of images _f2 and _f3 for frame difference analysis, and the calculation method is as follows.

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="T"\; filenames="HDA00002029130300021.png"\; state="\{\{state\}\}">T</figure-callout></span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; otherwise</span>}\end{array}\right.$

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="T"\; filenames="HDA00002029130300021.png"\; state="\{\{state\}\}">T</figure-callout></span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; otherwise</span>}\end{array}\right.$

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 1</span>}& {\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; and</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; otherwise</span>}\end{array}\right.$

Through the difference between images, obtain the binarized frame difference image. Since the frame-difference binarization image usually obtains the outline of the object, there must be some isolated points, local micro-regions, holes or incoherence caused by noise when the moving target is extracted. In the image processing, the morphological processing method is used to perform an erosion process to remove local tiny areas and isolated points in the video image. Then two dilation processes are performed to smooth the edge of the foreground object, and finally a complete moving object is obtained.

Step 6, since the traffic monitoring background will slowly change with factors such as sunshine, the texture feature background image is slowly updated in real time according to the formula (11), namely

${\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; or</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; img</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}\\ {\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>& {\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; img</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 11</span>}<span\; class="notranslate">\; )</span>$

Among them, a is an update parameter, representing the speed of background change. B _crs (t, x, y) is the value of pixel (x, y) in the texture background image of frame t, and F _crs (t, x, y) is the neighborhood image block of pixel (x, y) in the current frame image Roughness value, B _crs (t+1,x,y) is the updated texture background image. If the f(x, y) of the large area is 1, it means that there are moving vehicles in the lane, and the vehicles in the lane are not in a queuing state, and the background needs to be updated in real time. If F _img (x, y)=1, it means that the vehicles in the lane have formed a queuing queue, and the background cannot be updated, so as to prevent the foreground image from being publicized (11) to be updated as the background when the vehicle is stationary for a long time, and when the vehicle starts to move, a frame is generated difference, that is, after f(x, y) becomes 1, the background image is updated in real time.

(3) Measurement of vehicle queuing length

The tail position (x' _n , y' _n ) of the vehicle queue is obtained by image processing of the traffic monitoring video, and the angle α between the lane boundary line and the horizontal axis of the world coordinate system and the rotation angle φ of the camera are obtained by camera calibration.

First, solve the rotation angle φ of the camera so that it satisfies the conversion relationship between the image coordinate system and the world coordinate system established in the present invention. The coordinates (x _n , y _n ) of pixel n according to the rotation angle φ are

$\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}\\ \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}\end{array}\right]\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\end{array}\right]$

Secondly, convert (x _n , y _n ) of the coordinates of pixel n in the image coordinate system to (X _n , Y _n ) of the corresponding point N in the world coordinate system, and the conversion formula is as follows:

${\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}}<span\; class="notranslate">\; \&Center\; Dot;</span>\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}$

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}$

本发明使用数字摄像机获取交通监视场景的视频图像，结合摄像机快速标定方法和数字图像处理技术检测出交通路口车辆排队长度，达到实时交通流参数的采集。Finally, calculate the vehicle queue length

The invention uses a digital camera to acquire video images of traffic monitoring scenes, and combines the camera quick calibration method and digital image processing technology to detect the vehicle queuing length at traffic intersections to achieve real-time traffic flow parameter collection.

As shown in Figure 1, the present invention uses a PTZ video camera to acquire video images of road traffic monitoring scenes, sets the ROI of the road surface, uses texture features to detect the queuing status of vehicles in the ROI on the background image, and establishes an image through the rapid calibration of the camera The corresponding conversion relationship between the coordinate system and the world coordinate system, convert the detected pixel coordinates of the tail of the queue into world coordinate values, and calculate the length of the vehicle queue. PTZ cameras, namely Pan (translation), Tilt (tilt), Zoom (zoom), cameras that can change the viewing angle and zoom horizontally and vertically, are used in video surveillance systems.

In summary, the implementation process of the present invention is briefly summarized as the following steps:

1) Through the PTZ video camera installed at the traffic intersection, real-time acquisition of road condition video image information;

2) From the image information acquired by the video camera, select traffic markings with specific geometric features and size data, as shown in Figure 2. In the traffic markings on the road surface, select three collinear endpoints A, B, and C on the dotted line of the lane boundary, and obtain the distance between these three endpoints, the distance between the endpoints AB is h ₁ , and the distance between the endpoints AC is h ₂ . Take a certain point D on the vertical line of the dotted lane dividing line through a certain end point, and measure the distance d ₁ between this point and the intersection point.

According to the schematic diagram of the coordinate system shown in Figure 3, the coordinates of the above four endpoints in the image are (x' _a , y' _a ), (x' _b , y' _b ), (x' _c , y' _c ) and (x' _d , y' _d ).

When the camera is not installed completely horizontally, there is a rotation angle between the bottom edge of the CCD sensor and the horizontal line. Assuming that the rotation angle is φ, the four pixel coordinates of a, b, c and d are first rotated, and there is

$\left[\begin{array}{c}{x}_i\\ {\mathrm{the\; y}}_i\end{array}\right]=\left[\begin{array}{cc}\cos \mathrm{\&phi;}& -\sin \mathrm{\&phi;}\\ \sin \mathrm{\&phi;}& \cos \mathrm{\&phi;}\end{array}\right]\left[\begin{array}{c}{x}_i^{\&prime;}\\ {\mathrm{the\; y}}_i^{\&prime;}\end{array}\right]$ (where i=a,b,c,d) (1)

Calculate the intermediate variable $Y_A^{\&prime;}=\frac{h_1{h}_2({\mathrm{the\; y}}_b-{\mathrm{the\; y}}_c)}{({\mathrm{the\; y}}_a-{\mathrm{the\; y}}_b)h_2-({\mathrm{the\; y}}_a-{\mathrm{the\; y}}_c)h_1}---\left(2\right)$

$\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

In the world coordinate system, the auxiliary point A' as shown in Figure 4, the corresponding pixel point in the image is a', then the abscissa of the pixel point is a' is

${\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

In the image, a horizontal line is drawn through point d, and the point of intersection with the straight line where a, b, and c are located is point e. The pixel coordinates of point e are (x _e , y _e ), the slope of the known straight line L ₁ is k, and the intercept is b, as shown in Fig. 5 . The abscissa x _e of point e in the image plane is calculated as

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; e</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; b</span>}}{\mathrm{<span\; class="notranslate">\; k</span>}}$

then based on $\sin 2\mathrm{\&alpha;}=\frac{\mathrm{\&lambda;}-{\mathrm{the\; y}}_a}{\mathrm{\&lambda;}-{\mathrm{the\; y}}_c}\frac{{2d}_1(x_c-x_{a^{\&prime;}})}{h_2|x_e-x_a|},$ Calculate the angle α between the lane dividing line and the horizontal axis of the world coordinate system.

Convert the corresponding pixel coordinates of the calibration reference point D in the image into the coordinates of the world coordinate system, and the calculation formula is as follows.

${\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}<span\; class="notranslate">\; \&Center\; Dot;</span>\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

Assuming that the straight line passing through point D and straight line L ₁ intersect perpendicularly at point B, similarly convert the coordinates of pixel b into the coordinates of the world coordinate system, and calculate the distance |DB|, that is, the distance from point D to A, B, and C The distance of the point on the straight line L ₁ .

In the process of solving φ, repeat the formula (1) to formula (7) in the interval of φ∈[-15°,15°] to obtain the value of |DB| again. Use the Levenberg-Marquardt algorithm to search in the interval φ∈[-15°,15°], so that

φ=argmin(|DB|) (8)

(2) Video image analysis of vehicle queuing status

First, set the vehicle queuing detection area in the image, that is, the domain of interest ROI, and select the detection area from the traffic stop line in the lane to the predetermined lane detection area as the vehicle queuing detection area. As shown in Figure 6.

Secondly, the Tamura texture value of the neighborhood image block where the pixel (x, y) is located is used to measure the roughness of the point image, which is recorded as F _crs (x, y), and the neighborhood roughness value of each pixel F _crs ( x, y) form a new image as the texture feature background image B _crs .

Then, because the traffic monitoring background will slowly change with factors such as sunshine, for example, the texture background image in the daytime is obviously different from that in the nighttime, so the texture feature background image is slowly updated in real time according to formula (11). but

${\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; or</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; img</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}\\ {\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; crs</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>& {\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; img</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 11</span>}<span\; class="notranslate">\; )</span>$

Gaussian model analysis is used to extract the foreground image F _img from the background image.

Through the morphological processing method, a smooth and complete foreground image is obtained, that is, a binarized image of vehicles queuing.

Finally, from the lane stop line to the direction of lane queuing, the algorithm of moving detection window is used to search for the tail position of the queue, and the pixel coordinates N(x' _n , y' _n ) of the queue tail are obtained, as shown in Figure 6.

(3) Measurement of vehicle queuing length

After obtaining the precise φ angle, the position n(x' _n , y' _n ) of the vehicle queuing tail is obtained through image processing, and the pixel point n is first rotated and corrected according to the formula (1). The coordinates (x _n , y _n ) of pixel n according to the rotation angle φ are

$\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}\\ \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&phi;</span>}\end{array}\right]\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\end{array}\right]$

Secondly, convert (x _n , y _n ) of the coordinates of pixel n in the image coordinate system to (X _n , Y _n ) of the corresponding point N in the world coordinate system, and the conversion formula is as follows:

${\mathrm{<span\; class="notranslate">\; Y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}}<span\; class="notranslate">\; \&Center\; Dot;</span>\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}$

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{{\mathrm{<span\; class="notranslate">\; a</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}$

Finally, calculate the vehicle queue length, $L=\sqrt{{\left(x_{\mathrm{no}}\right)}^2+{\left(Y_{\mathrm{no}}\right)}^2}.$

Claims (6)

1. vehicle queue length measuring method based on the Pan/Tilt/Zoom camera Fast Calibration may further comprise the steps:

1) video image of usefulness PTZ video camera Real-time Obtaining traffic monitoring scene;

2) two traffic marking choosing square crossing in image consist of T-shaped camera calibration reference substance, obtain the pixel coordinate of four end points in the T-shaped scaling reference, measure the wherein between any two distance of three points of conllinear, and measure the distance that a bit arrives in addition this straight line;

3) set up model and the corresponding relation thereof of image coordinate system and world coordinate system, solve the inside and outside parameter of video camera, set up the conversion relation between the coordinate of the coordinate of image pixel and world coordinate system road surface corresponding point;

4) in image, choose stop line from the track to the lane detection district that presets as vehicle queue information detection zone, i.e. interest domain ROI, follow-up processing procedure is limited in the ROI; For any pixel (x, y), adopt the textural characteristics of its place neighborhood image piece to weigh the degree of roughness of this pixel, be designated as F _Crs(x, y) is with F _Crs(x, y) is as the pixel gray-scale value of grain background image; Go through all over the neighborhood roughness value F that calculates each pixel _Crs(x, y) forms the background image B that a width of cloth characterizes textural characteristics _Crs, and background image with sunshine light change can adaptive updates;

5) textural characteristics of realtime graphic and the textural characteristics of background image are compared, extract foreground image, after processing, morphology obtains the smooth of the edge, complete imperforate binaryzation foreground image, be vehicle queue information, from the stop line of track, search for to track queuing direction, adopt the algorithm search formation tail position of mobile detection window, obtain the image coordinate of tail of the queue pixel N;

6) the coordinate system conversion relation of setting up according to camera calibration is scaled the road surface corresponding point at the coordinate figure of world coordinate system, the queue length of vehicle queue when calculating traffic jam with the image coordinate of tail of the queue pixel N.

2. the camera calibration reference substance the vehicle queue length measuring method based on the Pan/Tilt/Zoom camera Fast Calibration according to claim 1, described step 2) is chosen track boundary dotted line L ₁The A of upper conllinear, B, 3 end points of C, and known these three end points distance between any two, terminal A B spacing is h ₁, terminal A C spacing is h ₂, the line L of outer another D of certain end points among A, B, the C and line ₂Perpendicular to L ₁, the D point is to straight line L ₁Distance be d ₁And known A, B, C, the D coordinate in image are respectively a(x ^' _a, y ^' _a), b(x ' _b, y ^' _b), c(x ^' _c, y ^' _c) and d(x ^' _d, y ^' _d).

3. the vehicle queue length measuring method based on the Pan/Tilt/Zoom camera Fast Calibration according to claim 1, the image coordinate system in the described step 3) and the model of world coordinate system and corresponding relation thereof, world coordinate system X _W-Y _W: X _WAxle was defined as the cam lens center and was parallel to the intersection on plane and the ground of ccd sensor, Y _WAxle was defined as camera optical axis and perpendicular to the intersection on plane and the ground on ground, Y _WThe axle positive dirction is along camera directed forward, X _WThe axle positive dirction is that level is pointed to right-hand; Its initial point O _WBe defined as X _WAxle and Y _WThe point of crossing of axle; Image coordinate system XOY: in image, establish image and be of a size of M * N, initial point is defined as image geometry center [(M-1)/2, (N-1)/2], and the X-axis level is pointed to left in the image, and Y-axis is vertically pointed to the below; When video camera was overlooked in the traffic scene scaling reference, the angle on optical axis and plane, road was θ, and the angle of lane line and world coordinate system is α, and the rotation angle of its optical axis of camera intrinsic is φ.

4. the vehicle queue length measuring method based on the Pan/Tilt/Zoom camera Fast Calibration according to claim 1, the camera calibration step in the described step 3) is:

When the non-complete level of video camera was installed, there were the anglec of rotation in bottom and the horizontal line of ccd sensor, suppose that rotation angle is φ, then at first a, b, c and four pixel coordinates of d are rotated, and have

(wherein i=a, b, c, d) (1)

And ask intermediate variable

The horizontal ordinate of auxiliary point a ' in the computed image

In image, cross the d point and make horizontal line, with a, b, c place straight line L ₁Intersection point be the e point, e point pixel coordinate is (x _e, y _e), the horizontal ordinate x that e is ordered in the plane of delineation _eFor

Foundation again

, obtain the angle α of lane line and world coordinate system transverse axis;

Be the coordinate of world coordinate system with the coordinate transformation of pixel d in the image, computing formula is as follows,

Suppose in world coordinate system, to cross straight line and the straight line L that D is ordered ₁Vertically meet at the B point, then the coordinate of corresponding point pixel b can be scaled according to formula (6), (7) coordinate of world coordinate system in image, calculates | and the distance of DB| is the D point to A, B, 3 places of C straight line L ₁Distance;

Adopt in Levenberg-Marquardt algorithm search φ ∈ [15 °, the 15 °] interval, so that

。

5. the vehicle queue length measuring method based on the Pan/Tilt/Zoom camera Fast Calibration according to claim 1, described step 4) uses the textural characteristics value of each neighborhood of pixels image block to set up the grain background image, and namely the gray-scale value of certain pixel is the roughness value of this neighborhood of pixels image block in the grain background image.

6. the vehicle queue length measuring method based on the Pan/Tilt/Zoom camera Fast Calibration according to claim 1, the step of described step 6) is:

After obtaining rotationangleφ, obtain to get vehicle queue tail of the queue position n (x through image processing and range searching ^' _n, y ^' _n), at first according to formula (9) pixel n is rotated rectification, the pixel n (x after the rectification _n, y _n) be

Secondly, with (the x of the coordinate of pixel n in the image coordinate system _n, y _n) be scaled (X of world coordinate system corresponding point N _n, Y _n), reduction formula is as follows:

At last, calculate vehicle queue length,

Images