JP3706018B2 - Moving object detection method, recording medium recording the program, and moving object detection apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing technique for detecting a moving object that moves at high speed, and more particularly to a moving object stably from a time-series image acquired by a line sensor camera, regardless of the size of the moving object to be detected. those related to the moving object detection how that made it possible to detect.
[0002]
[Prior art]
Conventionally, regarding a high-speed moving object detection device using a line sensor camera, an outdoor vehicle speed detection device (Reference 1: Japanese Patent Laid-Open No. 6-66820 “Speed measurement device”) has been proposed. Further, regarding a method of correcting the brightness of a line sensor camera and detecting a moving object when a detection target is determined indoors, a brightness correction method based on the speed of the moving object (Reference 2: JP-A-6-66633). Japanese Laid-open Patent Publication No. “Lightness Correction Device”) and a brightness correction method for detecting cracks on the surface of the rolled material (Reference 3: Japanese Patent Laid-Open No. 8-61929 “Rolled Material Shape Recognition Device”) have been proposed.
[0003]
In addition, as a method for detecting a change in the surface of a moving object using an area sensor, a surface defect inspection apparatus for inspecting a car body for scratches or irregularities (Reference 4: Japanese Patent Laid-Open No. 9-318337, “Surface Defects”). Inspection equipment ") has been proposed.
[0004]
[Problems to be solved by the invention]
When considering any high-speed moving object using a line sensor camera, the conventional high-speed moving object detection device does not perform correction processing for illumination fluctuations. Adjustment must be made, and brightness adjustment during acquisition is not possible.
[0005]
Moreover, the moving object detection method when the detection target is limited indoors can also be applied when the moving object moves at high speed. However, in the method of Document 1, the speed of the moving object must be measured in advance in order to control the charge time of a CCD (Charge Coupled Device) and correct the brightness according to the speed of the moving object. Reference 2 above describes the aperture adjustment and gain adjustment methods. However, since it aims to detect minute scratches on the rolled plate, the brightness is corrected by mistake for moving objects that cause large image changes. Therefore, high-speed moving objects cannot be detected.
[0006]
Further, the surface defect inspection apparatus using the area sensor described in Document 4 uses a method of correcting a luminance change due to illumination deterioration by adjusting the aperture. However, this method automatically adjusts the aperture of the target object before the inspection, and is not an aperture adjustment that can correct illumination variation at any time during the inspection. In the case of an area sensor, since the imaging cycle is long (in the case of NTSC, imaging is performed at 1/30 second intervals), there is no guarantee that the moving object can be detected stably when the speed of the moving object exceeds the imaging cycle.
[0007]
The present invention proposes a method for stably detecting a high-speed moving object by using a line sensor without depending on the size of the high-speed moving object to be detected. The present invention provides a technique that enables detection of a high-speed moving object without being restricted by fluctuations.
[0008]
[Means for Solving the Problems]
In the present invention, when a moving object moving along a predetermined route is photographed and detected by a line sensor camera installed at a fixed observation point, an average including line images for a plurality of immediately preceding cycles is firstly detected. If the brightness value is calculated and the average brightness value is outside the specified range, aperture adjustment or gain adjustment is performed, and the image change amount between the second acquired line image and the previous line image is calculated. , Detecting an image change point where the image change amount is larger than the threshold value, and thirdly, an aperture change amount or gain change amount within one period within a range where the image change amount accompanying the aperture adjustment or gain adjustment does not exceed the threshold value The main feature is to control .
[0009]
Specifically, the moving object detection device of the present invention calculates a mean luminance value of a line image group including a line image for a plurality of immediately preceding cycles, and means for acquiring a line image by photographing at a constant cycle interval, If the average luminance value is outside the specified range, means for adjusting the luminance value of the line image after the next period, means for determining an image change amount between each line image and the line image before each period, and the image change have a means for sensing the detected moving object image change point at which the amount is equal to or greater than a threshold, means for adjusting the luminance values of the line image, adjusts stop with respect to the amount of light entering the line sensor camera, the diaphragm It is means for controlling the amount of aperture change within one period within a range in which the image change amount accompanying the adjustment does not exceed the threshold value, or the gain is adjusted with respect to the output signal of the line sensor camera, and the image accompanying the gain adjustment The amount of change Characterized in that in the example no range is a means for controlling the gain variation within one period.
[0010]
The operation of the present invention is as follows. In the present invention, aperture adjustment, the image variation amount due to the gain adjustment does not exceed the threshold, without moving objects to be detected object to adjust the brightness value wrong even if would alter the entire background, moving objects Can be detected stably. In addition, since the change in the brightness value due to the brightness value adjustment of the present invention does not affect the threshold processing at the time of object detection, it is not necessary to perform detailed brightness value adjustment in advance. Moreover, since the constraint of the illumination variation is not stable time-series images can be acquired, so the pattern recognition processing can be easily performed with respect to the time-series image including a detected moving objects.
[0011]
A program for realizing the above processing means by a computer can be stored in an appropriate recording medium such as a portable medium memory, a semiconductor memory, or a hard disk that can be read by the computer.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings. In this embodiment, the high-speed moving object is a moving vehicle. The line sensor camera shall be installed on the upper part of the road and photograph the road.
[0013]
FIG. 1 is a diagram illustrating a configuration example of a high-speed moving object detection device according to an embodiment of the present invention. A storage device 103, a display device 104, and an image input device 105 are connected to the information processing apparatus 101 via a bus line 102. A threshold storage unit 106 and an image storage unit 107 are provided in the storage device 103. In the image input device 105, a line sensor camera 108, a diaphragm adjusting device 109, and a gain adjusting device 110 are provided.
[0014]
FIG. 2 shows an example of a measurement system of the high-speed moving object detection device according to the embodiment of the present invention. In FIG. 2, a high-speed moving object to be detected is a moving vehicle 201, and the line sensor camera 108 is installed so as to photograph the moving vehicle 201 from above.
[0015]
FIG. 3 is a flowchart of the high-speed moving object detection apparatus. In FIG. 3, 301 is a processing step for acquiring a line image from the line sensor camera 108, 302 is a processing step for calculating an image change amount between the acquired line image and the line image one cycle before, and 303 is a threshold value for the image change amount. A processing step for determining whether the image is larger, 304 a processing step for acquiring a line image of a specified number of lines, 305 a processing step for storing an image of a moving image, and 307 a line surface image for a plurality of previous cycles. 308 is a processing step for calculating the included average luminance value, 308 is a processing step for determining whether the average luminance value is outside the specified range, and 309 is a processing step for performing aperture adjustment and gain adjustment.
[0016]
An example of the high-speed moving object device will be described with reference to the flowchart of FIG. First, the case where the average luminance value of the acquired line image is within a predetermined range will be described. First, a line image is acquired from the line sensor camera 108 (step 301 in FIG. 3).
[0017]
FIG. 4 shows a line image acquired from the line sensor camera 108 and a graph of the temporal image change amount. Since the line sensor camera 108 is fixed and images the same part at a constant cycle, when the moving vehicle 201 passes on the imaging line, an image of the moving vehicle 201 can be captured like a line image 401 of the line sensor camera.
[0018]
Next, an image change amount between the acquired line image and the previous line image is calculated (step 302). Here, the image change amount D is the average value D of luminance absolute difference values for each pixel of the line image D = Σ _{i = 1} ^N | L _i −L ′ _i | / N (1)
It was. Here, L _i represents the i-th pixel value of the acquired line image, L ′ _i represents the i-th pixel value of the line image one cycle before, and N represents the number of pixels of the line image. At this time, the temporal change of the image change amount with respect to the line image 401 of the line sensor camera of FIG. 4 becomes like the temporal change 402 of the image change amount between the line image and the line image one cycle before. Here, the image change amount is the average value of the luminance absolute difference values, but the mean square error D2 = Σ _{i = 1} ^N (L _i −L ′ _i ) ² / N (2)
Also, the correlation value C = Σ _{i = 1} ^N (L _i −L _mean ) × (L ′ _i −L ′ _mean ) / (σ _L × σ _{L '} ) ... (3)
It is good. Here, L _mean , L ′ _mean and σ _L , σ _{L ′} represent the average value and the standard deviation. C has a value from -1 to 1, and the closer to 1, the smaller the image change amount. If the difference between images can be expressed numerically by other methods, the value may be treated as an image change amount.
[0019]
Next, it is determined whether the image change amount is larger than a threshold value (step 303). Since the image change amount of the line image increases when the moving vehicle 201 passes the shooting line, the image change point 403 can be detected by threshold processing for the image change amount using the threshold 404 for the image change amount.
[0020]
Next, a line image having the specified number of lines is acquired (step 304). Since the image change point, that is, the tip of the moving vehicle 201 can be detected in step 303, line images for a predetermined time are acquired using this detection time as a start time. Thus, by acquiring a line image, an image of a high-speed moving object can be stored.
[0021]
Next, a method for adjusting the brightness value of the line image to be within the specified range when the average brightness value of the line image is outside the specified range will be described. First, a line image is acquired from the line sensor camera 108 (step 301), and an average luminance value including line images for a plurality of immediately preceding cycles is calculated (step 307).
[0022]
As for the number of the plurality of lines, it is desirable that the ratio of the moving vehicle in the two-dimensional image composed of the plurality of lines is 1/2 or less. This is to suppress unnecessary adjustment by the moving vehicle. The specific number of lines is determined from the shooting time interval and the expected speed of the moving vehicle.
[0023]
FIG. 5 shows a graph representing a temporal change with respect to the average luminance value of the line images for a plurality of periods. Here, the vertical axis represents the average luminance value 501, and the horizontal axis represents time 502. The specified range 503 indicates a range in which a high-speed moving object can be detected correctly by threshold processing for the image change amount. Time A504 is when the average luminance value is larger than the specified range, and time B505 is when the average luminance value is smaller than the specified range.
[0024]
After calculating the average luminance value including the line images for the immediately preceding plural cycles, it is determined whether the average luminance value is outside the specified range (step 308). FIG. 6 is a diagram showing a temporal change in the image change amount when the moving vehicle 201 passes at time A504 in the case where the average luminance value is larger than the specified range. Similarly, FIG. 7 is a diagram showing a temporal change in the image change amount when the moving vehicle 201 passes at time B505 when the average luminance value is smaller than the specified range. As described above, when the average luminance value is outside the specified range, even if an image change occurs due to saturation or shortage of the luminance value, the image change amount does not satisfy the threshold value, and the moving vehicle 201 cannot be detected. .
[0025]
Therefore, when the average luminance value is outside the specified range, aperture adjustment and gain adjustment are performed (step 309). Specifically, aperture adjustment is performed by adjusting the aperture of a lens attached to the line sensor camera 108, and gain adjustment is performed by adjusting the gain (gain) of the output signal of the CCD camera.
[0026]
FIG. 8 shows temporal changes in the amount of image change due to aperture adjustment and gain adjustment. At this time, if the image change amount due to the adjustment is smaller than the threshold value 803 for the image change amount, the average luminance value 501 can be adjusted to fall within the specified range 503 without being detected as the moving vehicle 201. Thus, even if there is a lighting fluctuation such as sunlight, the moving vehicle 201 can be detected by the processing of steps 301 to 305 by always adjusting the average luminance value 501 to the specified range 503.
[0027]
If the moving object is slow, it can be detected by increasing the shooting cycle or increasing the number of lines when obtaining the average luminance value during luminance adjustment.
[0028]
【The invention's effect】
As described above, according to the present invention, even when an image changes greatly due to a high-speed moving object in an environment where illumination fluctuations occur, a high-speed moving object is detected by a line sensor camera installed at a fixed observation point. Can be captured and detected.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a configuration example of a high-speed moving object detection device according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating an example of a measurement system of a high-speed moving object detection device.
FIG. 3 is a flowchart of the high-speed moving object detection device.
FIG. 4 is a diagram illustrating an example of a line image acquired from a line sensor camera and its temporal image change amount.
FIG. 5 is a diagram illustrating an example of a temporal change with respect to an average luminance value of a line image for a plurality of periods.
FIG. 6 is a diagram illustrating an example of a temporal change in an image change amount when a moving vehicle passes when an average luminance value is larger than a specified range.
FIG. 7 is a diagram illustrating an example of a temporal change in an image change amount when a moving vehicle passes when an average luminance value is smaller than a specified range.
FIG. 8 is a diagram illustrating an example of a temporal change in an image change amount due to aperture adjustment and gain adjustment.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 101 Information processing apparatus 102 Bus line 103 Memory | storage device 104 Display apparatus 105 Image input apparatus 106 Threshold storage part 107 Image storage part 108 Line sensor camera 109 Aperture adjustment apparatus 110 Gain adjustment apparatus

Claims (6)

A method of photographing and detecting a moving object moving along a predetermined route with a line sensor camera installed at a fixed observation point,
A procedure for capturing line images by shooting at regular intervals;
Calculating the average luminance value of the line image group including the line images for the immediately preceding multiple periods, and adjusting the luminance value of the line image after the next period if the average luminance value is outside the specified range;
A procedure for determining the amount of image change between each line image and the line image one cycle before;
Wherein the image variation amount is Yes and the procedure for detecting a moving object by detecting an image change point which is a predetermined threshold value or more,
In the procedure for adjusting the luminance value of the line image, aperture adjustment is performed with respect to the amount of light entering the line sensor camera, and the aperture change within one cycle is within a range where the image change amount accompanying the aperture adjustment does not exceed the threshold value. A moving object detection method characterized by controlling an amount . A method of photographing and detecting a moving object moving along a predetermined route with a line sensor camera installed at a fixed observation point,
A procedure for capturing line images by shooting at regular intervals;
Calculating the average luminance value of the line image group including the line images for the immediately preceding multiple periods, and adjusting the luminance value of the line image after the next period if the average luminance value is outside the specified range;
A procedure for determining the amount of image change between each line image and the line image one cycle before;
Detecting a moving object by detecting an image change point at which the image change amount is a predetermined threshold value or more, and
In the procedure of adjusting the luminance value of the line image, the gain is adjusted with respect to the output signal of the line sensor camera, and the gain change within one cycle is within a range where the image change amount accompanying the gain adjustment does not exceed the threshold value. A moving object detection method characterized by controlling an amount . In the moving object detection method of Claim 1 or Claim 2 ,
A method of detecting a moving object, comprising: a step of acquiring a line image for a predetermined period after detecting the moving object and storing a two-dimensional time-series image of the moving object. A recording medium on which a program for causing a computer to execute the moving object detection method according to any one of claims 1 to 3 is recorded. A device that captures and detects moving objects that move along a predetermined route using a line sensor camera installed at a fixed observation point.
Means for capturing line images by taking images at regular intervals;
Means for calculating an average luminance value of a line image group including line images for a plurality of immediately preceding cycles, and adjusting the luminance value of the line image after the next cycle if the average luminance value is out of a specified range;
Means for determining an image change amount between each line image and the line image one cycle before;
Means for detecting a moving object by detecting an image change point at which the image change amount is equal to or greater than a predetermined threshold ,
Further, the means for adjusting the luminance value of the line image adjusts the aperture with respect to the amount of light entering the line sensor camera, and changes the aperture within one cycle within a range where the image change amount accompanying the aperture adjustment does not exceed the threshold value. A moving object detection device characterized by being a means for controlling the amount . A device that captures and detects moving objects that move along a predetermined route using a line sensor camera installed at a fixed observation point.
Means for capturing line images by taking images at regular intervals;
Means for calculating an average luminance value of a line image group including line images for a plurality of immediately preceding cycles, and adjusting the luminance value of the line image after the next cycle if the average luminance value is out of a specified range;
Means for determining an image change amount between each line image and the line image one cycle before;
Means for detecting a moving object by detecting an image change point at which the image change amount is equal to or greater than a predetermined threshold,
In addition, the means for adjusting the luminance value of the line image adjusts the gain with respect to the output signal of the line sensor camera, and the gain change within one cycle within a range where the image change amount accompanying the gain adjustment does not exceed the threshold value. A moving object detection device characterized by being a means for controlling the amount .

Images