CN114666550B - Device and method for detecting contact point parameters of close-fit inspector based on image processing - Google Patents

Abstract

The invention discloses a device and a method for detecting contact point parameters of a close-fitting inspector based on image processing, wherein the device comprises a camera, a light supplementing lamp, a communication line, a power supply, a processor and a display; shooting the moving state of the movable contact of the close-fitting inspector by a camera; the processor collects the start and end state images of the close-fitting inspector to be processed, splices the images and selects the area to be processed on the close-fitting inspector image; filtering the interference pixel points of the area to be processed to obtain an edge image containing the outline of the movable contact of the close-fitting inspector; fitting, positioning and correcting distortion of pixel points on the edge image to obtain a movable contact point position image; and processing the obtained moving contact position image, and positioning the center coordinates and the diameter of the moving contact moving ring through extracting the characteristics of the moving contact to obtain the contact point parameters of the close-fitting inspector. The invention has simple structure, convenient installation and high protective performance, realizes the functions of real-time monitoring and alarming of the close-fitting inspector, and can intuitively see the working condition of the close-fitting inspector.

Description

Image processing-based contact point parameter detection device and method for close contact inspection device

Technical Field

The invention relates to the field of parameter detection of a turnout close contact checker, and in particular to a device and method for detecting contact point parameters of a close contact checker based on image processing.

Background Art

The close fit checker is used to check the close fit of the point rail and the heart rail, and can also be used to indicate the cut-off when the switch is squeezed. The most popular close fit checker on the market is the JM-A type close fit checker, which is suitable for turnouts with a straight-through train speed of more than 120km/h. It is used to detect whether the gap between two traction points is greater than 5mm. However, it can only check the close fit of one point rail, so two close fit checkers are required for each set of two point rails of the turnout, installed on both sides of the line.

As the close-fitting checker is widely used across the country, some problems have been exposed with the product itself. For example, the close-fitting checker can only check the close-fitting and repulsion status through manual on-site inspections. The monitoring efficiency is low, the speed is slow, the monitoring process is cumbersome, and it is impossible to provide early warning.

Summary of the invention

The purpose of the present invention is to provide a contact point parameter detection device and method for a close fit inspector based on image processing, which has a simple structure, is easy to install and has high protection performance. Machine vision is used instead of manual inspection to realize real-time monitoring and alarm functions of the close fit inspector, and the working status of the close fit inspector can be intuitively seen.

The technical solution to achieve the purpose of the present invention is: a contact point parameter detection device for a close contact checker based on image processing, comprising a first camera, a second camera, a first fill light, a second fill light, a communication line, a power supply, a processor, and a display, wherein:

The first camera and the second camera are fixed above the moving contact and moving ring of the close contact inspector, and are used to capture the moving process images of the moving contact and moving ring of the close contact inspector;

The first fill light and the second fill light are arranged between the first camera and the second camera, and are used to fill light for the moving contact and moving ring of the close-fitting inspector during shooting;

The communication line and the power supply are arranged on the side of the close contact inspector, the power supply is used to supply power to the first camera, the second camera, the first fill light, and the second fill light, and the communication line is used to transmit the image information captured by the first camera and the second camera;

The processor receives the image information captured by the first camera and the second camera, selects a processing area on the image, filters the interfering pixel points in the detection area, and obtains an edge image containing the contour of the moving contact point of the close-fitting inspector; fits and locates the pixel points on the edge image to obtain a moving contact point position image; and detects the contact point parameters by feature comparison;

The display is used to display the detection results of the contact point parameters.

A method for detecting contact point parameters of a close contact inspection device based on image processing, characterized in that the steps are as follows:

Step 1: Place the first camera and the second camera inside the close contact inspection box, and use a non-contact method to detect the contact points of the close contact inspection box;

Step 2: observe the internal situation of the close contact inspection device through the first camera and the second camera, shoot the motion state of the moving contact of the close contact inspection device, and store the initial and final state images;

Step 3, the processor collects the initial and final state images of the close contact inspector to be processed, splices the images and selects the area to be processed on the close contact inspector image;

Step 4, filtering the interfering pixels in the area to be processed to obtain an edge image containing the contour of the moving contact of the close-fitting inspector;

Step 5: Fit, locate and correct the distortion of the pixel points on the edge image to obtain a moving joint position image;

Step 6. Process the acquired moving contact position image of the close fit inspector, locate the center coordinates and diameter of the moving contact moving ring by extracting the features of the moving contact, and obtain the contact point parameters of the close fit inspector, including the contact group pressure of the close fit inspector, the depth of the moving ring of the close fit inspector connected to the static contact piece, and the displacement distance of the moving ring of the close fit inspector at the beginning and end of operation.

Compared with the prior art, the present invention has the following significant advantages: (1) The structure of the whole machine is simple and clear: the hardware part only includes power supply, communication line, camera and fill light; (2) By detecting the contact point of the close contact checker, pictures of the initial and final states of the moving contact of the close contact checker, the penetration depth of the moving contact ring and the pressure of the static contact piece at the moving contact are obtained, and the detection accuracy is high, which can realize the real-time monitoring, pre-alarm and alarm functions of the close contact checker; (3) Machine vision is used instead of manual labor. This contactless detection method has high detection efficiency, low maintenance cost after one-time investment cost, and is not affected by environmental and time factors. The working status of the close contact checker can be intuitively seen, and timely detection is convenient for predicting risks to ensure safe operation of the railway; (4) The close contact checker can be monitored in real time, which improves the monitoring efficiency and visualizes the processing results. At the same time, the monitoring data can be recorded to timely discover possible problems of the close contact checker, and early detection and maintenance can be carried out to ensure that high-speed EMUs can pass through the switch safely and smoothly.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flow chart of a method for detecting contact point parameters of a close contact checker based on image processing provided by an embodiment of the present invention.

FIG. 2 is a schematic structural diagram of a contact point detection device for a close contact checker based on image processing provided in an embodiment of the present invention.

FIG. 3 is a diagram showing the software interface of the contact point parameter detection device of the close contact checker of the present invention.

FIG. 4 is a flowchart of the working process of the contact point parameter detection device of the close contact checker in the present invention.

DETAILED DESCRIPTION

The present invention discloses a contact point parameter detection device for a close contact inspection device based on image processing, characterized in that it comprises a first camera 1, a second camera 2, a first fill light 3, a second fill light 4, a communication line 5, a power supply 6, a processor 7, and a display 8, wherein:

The first camera 1 and the second camera 2 are fixed above the moving contact and moving ring of the close contact inspector, and are used to capture the moving process images of the moving contact and moving ring of the close contact inspector;

The first fill light 3 and the second fill light 4 are arranged between the first camera 1 and the second camera 2, and are used to fill light for the moving contact and moving ring of the close-fitting inspector during shooting;

The communication line 5 and the power supply 6 are arranged on the side of the close contact inspector, the power supply 6 is used to supply power to the first camera 1, the second camera 2, the first fill light 3, and the second fill light 4, and the communication line 5 is used to transmit the image information captured by the first camera 1 and the second camera 2;

The processor 7 receives the image information captured by the first camera 1 and the second camera 2, selects a processing area on the image, filters the interfering pixels in the detection area, and obtains an edge image containing the contour of the moving contact of the close-fitting inspector; fits and locates the pixels on the edge image to obtain a moving contact position image; and detects the contact point parameters by feature comparison;

The display 8 is used to display the detection results of the contact point parameters.

Furthermore, the processor 7 selects a processing area on the image, filters the interfering pixels in the detection area, and obtains an edge image containing the contour of the moving contact of the close-fitting inspector, as follows:

Convert the area to be processed on the close contact checker image into a grayscale image in RGB format, and use the Canny operator to perform edge detection on the grayscale image to obtain a first edge image containing the contact points of the close contact checker; the area to be processed is the area where the contact point group ring exists;

A connected domain search is performed on the edges of the first edge image, and a connected domain with a number of pixels less than a set value is deleted to obtain a second edge image.

Furthermore, the processor 7 fits and locates the pixel points on the edge image to obtain the moving joint position image, including using Zhang Zhengyou's plane calibration method to calibrate the camera. During the camera calibration stage, a printed template is attached to a plane, and then multiple template images are taken from different angles. The feature points in the template image are detected and the distortion of the camera model is considered. The optimization algorithm is used to solve and obtain accurate internal and external parameters and distortion coefficients of the camera.

Furthermore, the processor 7 detects the contact point parameters by feature comparison, as follows:

First, calculate the gradient of each pixel in the image, then traverse the non-zero pixels in the image after edge detection, draw a straight line along the gradient of the non-zero pixels, and accumulate the points passed by the straight line. The larger the point in the accumulator, the more likely it is to be the center of the circle. The points in the accumulator that are greater than the given threshold and greater than all of their neighbors are retained as candidate points, and then these candidate points are sorted in descending order according to the accumulated value to ensure that the point most likely to be the center of the circle appears first; after selecting the candidate point of the center of the circle, consider all non-zero pixels for each center and make a radius for the possible center of the circle;

By setting relevant parameters of Hough circle detection, the edge of the circular ring of the moving contact of the close fitting inspector is located in the image; by extracting the center radius of the circle, the center position of the moving contact and the diameter of the moving ring are located; by detecting the diameter of the moving ring, the opening distance of the static joint is obtained, and the functional relationship between the opening distance and the force is brought into play to obtain the pressure of the contact group of the close fitting inspector; by comparing the position of the center of the moving ring of the close fitting inspector in the initial and final states, the pixel distance of the moving ring movement is obtained, and by camera calibration, the actual distance of the moving ring movement and the penetration depth of the moving ring are obtained; the detection of the close fitting inspector is transformed from artificial abstract observation to digital calibration.

Furthermore, the display 8 is used to display the detection results output by the processor 7, including images of the initial and final states of the moving contact movement, the driving depth of the moving contact ring, and the pressure of the static contact piece driven into the moving contact.

The present invention provides a method for detecting contact point parameters of a close contact inspection device based on image processing, and the steps are as follows:

Step 1, setting the first camera 1 and the second camera 2 inside the close contact inspection box, and adopting a non-contact method to detect the contact point of the close contact inspection;

Step 2: observe the internal situation of the close contact inspection box through the first camera 1 and the second camera 2, shoot the motion state of the moving contact of the close contact inspection device, and store the initial and final state images;

Step 3, the processor 7 collects the initial and final state images of the close contact inspection device to be processed, splices the images and selects the area to be processed on the close contact inspection device image;

Step 4, filtering the interfering pixels in the area to be processed to obtain an edge image containing the contour of the moving contact of the close-fitting inspector;

Step 5: Fit, locate and correct the distortion of the pixel points on the edge image to obtain a moving joint position image;

Step 6. Process the acquired moving contact position image of the close fit inspector, locate the center coordinates and diameter of the moving contact moving ring by extracting the features of the moving contact, and obtain the contact point parameters of the close fit inspector, including the contact group pressure of the close fit inspector, the depth of the moving ring of the close fit inspector connected to the static contact piece, and the displacement distance of the moving ring of the close fit inspector at the beginning and end of operation.

Furthermore, in step 4, the interfering pixels of the area to be processed are filtered to obtain an edge image containing the contour of the moving contact of the close-fitting inspector, as follows:

Convert the area to be processed on the close contact checker image into a grayscale image in RGB format, and use the Canny operator to perform edge detection on the grayscale image to obtain a first edge image containing the contact points of the close contact checker; the area to be processed is the area where the contact point group ring exists;

A connected domain search is performed on the edges of the first edge image, and a connected domain with a number of pixels less than a set value is deleted to obtain a second edge image.

Furthermore, step 5 describes fitting, positioning and distortion correction of pixel points on the edge image to obtain a moving joint position image, including calibrating the camera using Zhang Zhengyou's plane calibration method. During the camera calibration stage, a printed template is attached to a plane, and then multiple template images are taken from different angles. Feature points in the template image are detected and the distortion of the camera model is considered. An optimization algorithm is used to solve and obtain accurate internal and external parameters and distortion coefficients of the camera.

Furthermore, the moving contact position image of the close contact checker obtained by the processing in step 6 is extracted by extracting the features of the moving contact, locating the center coordinates and diameter of the moving ring of the moving contact, and obtaining the contact point parameters of the close contact checker, which are as follows:

First, calculate the gradient of each pixel in the image, then traverse the non-zero pixels in the image after edge detection, draw a straight line along the gradient of the non-zero pixels, and accumulate the points passed by the straight line. The larger the point in the accumulator, the more likely it is to be the center of the circle. The points in the accumulator that are greater than the given threshold and greater than all of their neighbors are retained as candidate points, and then these candidate points are sorted in descending order according to the accumulated value to ensure that the point most likely to be the center of the circle appears first; after selecting the candidate point of the center of the circle, consider all non-zero pixels for each center and make a radius for the possible center of the circle;

By setting relevant parameters of Hough circle detection, the edge of the circular ring of the moving contact of the close fitting inspector is located in the image; by extracting the center radius of the circle, the center position of the moving contact and the diameter of the moving ring are located; by detecting the diameter of the moving ring, the opening distance of the static joint is obtained, and the functional relationship between the opening distance and the force is brought into play to obtain the pressure of the contact group of the close fitting inspector; by comparing the position of the center of the moving ring of the close fitting inspector in the initial and final states, the pixel distance of the moving ring movement is obtained, and by camera calibration, the actual distance of the moving ring movement and the penetration depth of the moving ring are obtained; the detection of the close fitting inspector is transformed from artificial abstract observation to digital calibration.

The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

Example

In conjunction with FIG1 , this embodiment provides a method for detecting contact point parameters of a close contact checker based on image processing, comprising the following steps:

Step 1: collect the image of the close fitting detector to be processed, and select the area to be processed on the close fitting detector image;

There are many ways to collect images, such as conventional collection devices such as video cameras and cameras. After taking the bottle cap image through the collection device, it is uploaded to the processor for further processing.

Step 2: Filter the interfering pixels in the area to be processed to obtain an edge image containing the contour of the moving contact of the close-fitting inspector, specifically including:

The area to be processed on the close contact checker image is converted into a grayscale image in RGB format, and the Canny operator is used to perform edge detection on the grayscale image to obtain a first edge image containing the close contact checker contact points.

The area that needs to be processed is the area with dense pixels and the area where the contact group ring exists. The Canny operator is a method widely used in edge detection. Using the Canny operator to detect the edge of the grayscale image can obtain a more accurate bottle cap edge image.

A connected domain search is performed on the edges of the first edge image, and a connected domain with a number of pixels less than a set value is deleted to obtain a second edge image.

Step 3: Fitting, locating and correcting the pixel points on the edge image to obtain a close contact point image of the inspection device, including calibrating the camera using Zhang Zhengyou's plane calibration method, and the specific steps include:

During the camera calibration stage of the system, a template with high printing accuracy is attached to a plane, and then several template images are taken from different angles. The feature points in the image are detected and the distortion of the camera model is considered. The optimization algorithm is used to refine the accurate internal and external parameters and distortion coefficients of the camera.

The camera is calibrated using Zhang Zhengyou's plane calibration method. The basic principle is as follows:

Assume that the template plane is on the plane of the world coordinate system Z = 0, where K is the intrinsic parameter matrix of the camera, M = [X, Y, 1] ^T is the homogeneous coordinates of the points on the template plane, m = [u, v, 1] ^T is the homogeneous coordinates of the corresponding points projected from the points on the template plane to the image plane, [r ₁ , r ₂ , r ₃ ] and t are the rotation matrix and translation vector of the camera coordinate system relative to the world coordinate system, respectively;

According to the properties of the rotation matrix, and ||r ₁ ||＝||r ₂ ||＝1, each image can obtain the following two basic constraints on the internal parameter matrix:

Since the camera has 5 unknown intrinsic parameters, when the number of images captured is greater than or equal to 3, K can be solved linearly and uniquely;

During the camera calibration stage of the system, a template with high printing accuracy is attached to a plane, and then several template images are taken from different angles. The feature points in the image are detected and the distortion of the camera model is considered. The optimization algorithm is used to refine the accurate internal and external parameters and distortion coefficients of the camera.

Step 4: Detect the contact point parameters of the close contact detector by feature extraction, including:

Convert the image into Hough transform space, convert the points in the rectangular coordinate system into the polar coordinate system. By deriving mathematical relationships, converting the straight line detection problem in the image space into the Hough space becomes a problem of detecting the convergence points of the curve. The Hough gradient method is selected to solve the detection of the contact circle.

First, calculate the gradient of each pixel in the image, then traverse the non-zero pixels in the image after edge detection, draw a straight line along the gradient of the point, and accumulate the points passed by the line. The larger the point in the accumulator, the more likely it is the center of the circle. Keep the points in the accumulator that are greater than the given threshold and greater than all its neighbors, and then sort these points in descending order according to the accumulated value to ensure that the point most likely to be the center of the circle appears first. After selecting the candidate center point, consider all non-zero pixels for each center and make a radius for a possible center of the circle.

By setting the relevant parameters of the Hough circle detection, the edge of the ring of the dynamic ring of the close contact checker is located in the picture. By extracting the center radius of the circle, the center position of the dynamic contact ring and the diameter of the dynamic ring are located. By detecting the diameter of the dynamic ring, the opening distance of the static joint is obtained, and the functional relationship between the opening distance and the force is brought into the close contact checker contact group pressure; the pixel distance of the dynamic ring movement is obtained by the position of the center of the dynamic ring of the close contact checker at the initial and final states, and the actual distance of the dynamic ring movement and the dynamic ring penetration depth are obtained through camera calibration.

It should be noted that the thresholds and other related parameters set in the present invention vary in size according to different accuracy requirements and can be flexibly set according to actual needs.

This embodiment also provides a contact point parameter detection device for a close contact inspection device based on image processing, as shown in FIG2 , including a first camera 1, a second camera 2, a first fill light 3, a second fill light 4, a communication line 5, a power supply 6, a processor 7, and a display 8, wherein:

The first camera 1 and the second camera 2 are fixed above the moving contact and moving ring of the close contact inspector, and are used to capture the moving process images of the moving contact and moving ring of the close contact inspector;

The first fill light 3 and the second fill light 4 are arranged between the first camera 1 and the second camera 2, and are used to fill light for the moving contact and moving ring of the close-fitting inspector during shooting;

The communication line 5 and the power supply 6 are arranged on the side of the close contact inspector, the power supply 6 is used to supply power to the first camera 1, the second camera 2, the first fill light 3, and the second fill light 4, and the communication line 5 is used to transmit the image information captured by the first camera 1 and the second camera 2;

The processor 7 receives the image information captured by the first camera 1 and the second camera 2, selects a processing area on the image, filters the interfering pixels in the detection area, and obtains an edge image containing the contour of the moving contact of the close-fitting inspector; fits and locates the pixels on the edge image to obtain a moving contact position image; and detects the contact point parameters by feature comparison;

The display 8 is used to display the detection results of the contact point parameters.

As a specific example, the camera and fill light are fixed above the contact point of the close-fitting inspector, and the side of the close-fitting inspector box completes the packaging circuit part and the collection and transmission unit.

As a specific example, the camera is fixedly connected to the fixture, the strip fill light is fixed between the cameras, and the power supply is provided by a mobile power supply on the side.

As a specific example, the internal circuit elements are connected by connecting pieces and welding.

As a specific example, the opening distance of the static joint is obtained by detecting the diameter of the moving ring, and the functional relationship between the opening distance and the force is brought into play to obtain the pressure of the contact group of the close fitting inspector; the pixel distance of the moving ring movement is obtained by comparing the position of the center of the moving ring of the close fitting inspector in the initial and final states, and the actual distance of the moving ring movement and the penetration depth of the moving ring are obtained through conversion between the world coordinate system, the camera coordinate system, the image coordinate system and the pixel coordinate system.

As a specific example, it also includes PLC, acquisition card, and power supply, wherein: the PLC control system unit runs in coordination with the equipment; the acquisition card transmits the collected data to the host computer for real-time viewing of the detection results; and the mobile power supply supplies power to the entire device.

In conjunction with FIG. 3 and FIG. 4 , the working process of the contact point parameter detection device of the close contact inspection device based on image processing in this embodiment is as follows:

S1. Insert the USB interface of the communication cable into the host computer and turn on the power switch;

S2. Acquire images through software control. Two industrial cameras are set on the upper part of the dynamic ring of the close fit inspector to be inspected. The effective focal length of the camera is 2.96±0.03mm. When the photo button is clicked on the software interface, the camera is triggered to collect the image of the close fit inspector. Each camera takes half of the image of the close fit inspector.

S3. After the forward and reverse direction shooting is completed, click to end the shooting, perform image transmission processing, analyze the close contact inspection device through the functional algorithm, and analyze the coordinates and diameter parameters of the moving contact ring;

S4. Click Display Results to display pictures of the initial and final states of the moving contact movement after processing, the penetration depth of the moving contact ring, and the pressure of the static contact piece at the moving contact on the interface. The test results can be observed in real time, and an alert will be issued if any abnormal situation occurs.

S5. Click Save Record to record the test results.

In summary, the overall structure of the present invention is simple and clear: the hardware part only has power supply, communication line, camera, and fill light; by detecting the contact point of the close fit checker, pictures of the initial and final states of the moving contact of the close fit checker, the penetration depth of the moving contact ring, and the pressure of the static contact piece at the moving contact are obtained, and the detection accuracy is high, which can realize the real-time monitoring, pre-alarm, alarm and other functions of the close fit checker. In addition, the present invention also uses machine vision to replace manual labor. This contactless detection method has high detection efficiency, low maintenance cost after a one-time investment cost, and the detection is not affected by environmental and time factors. The working status of the close fit checker can be intuitively seen, and timely detection is convenient for predicting risks to ensure the safe operation of the railway. In short, the present invention can monitor the close fit checker in real time, improve the monitoring efficiency, and visualize the processing results. At the same time, it can record the monitoring data, timely discover possible problems with the close fit checker, and repair them early to ensure that the high-speed EMU can pass the switch safely and smoothly.

Claims (7)

1. The utility model provides a close-fitting inspector contact point parameter detection device based on image processing, its characterized in that, including first camera (1), second camera (2), first light filling lamp (3), second light filling lamp (4), communication line (5), power (6), processor (7), display (8), wherein:

the first camera (1) and the second camera (2) are fixed above the movable contact moving ring of the close-fitting inspector and are used for shooting moving process images of the movable contact moving ring of the close-fitting inspector;

The first light supplementing lamp (3) and the second light supplementing lamp (4) are arranged between the first camera (1) and the second camera (2) and are used for supplementing light for the movable contact movable ring of the close-fitting inspector in the shooting process;

the communication line (5) and the power supply (6) are arranged on the side surface of the close-fitting inspector, the power supply (6) is used for supplying power to the first camera (1), the second camera (2), the first light supplementing lamp (3) and the second light supplementing lamp (4), and the communication line (5) is used for transmitting image information shot by the first camera (1) and the second camera (2);

The processor (7) receives image information shot by the first camera (1) and the second camera (2), selects a processing area on the image, filters interference pixel points of the detection area, and obtains an edge image containing the outline of the movable contact of the close-fitting inspector; fitting and positioning pixel points on the edge image to obtain a movable contact point position image; by means of feature comparison, the parameters of the contact are detected, and the method specifically comprises the following steps:

Firstly calculating the gradient of each pixel point in an image, traversing non-0 pixel points in the image after edge detection, drawing a straight line along the gradient of the non-0 pixel points, accumulating points through which the straight line passes, accumulating points with larger points in an accumulator being more likely to be circle centers, reserving points which are larger than a given threshold value and are larger than all neighbors of the points in the accumulator as candidate points, and arranging the candidate points in descending order according to accumulated values so as to ensure that the point most likely to be the circle center appears earliest; after selecting circle center candidate points, considering all non-0 pixels for each center, and making a radius for a possible circle center;

Positioning the edge of a movable joint moving ring of the close-fitting inspector in the image through the relevant parameter setting of Hough circle detection; the circle center position and the diameter of the movable joint movable ring are positioned through the extraction of the circle center radius; acquiring the opening distance of the static contact sheet by detecting the diameter of the movable ring, and carrying in the functional relation between the opening distance and the stress to obtain the contact group pressure of the close-fitting inspector; obtaining the pixel distance of the moving ring by closely attaching the position of the circle center of the moving ring of the inspector to the starting and ending states, and obtaining the actual distance of the moving ring and the driving depth of the moving ring by calibrating a camera; converting the detection of the close-fitting inspector from manual abstract observation into digital calibration;

and the display (8) is used for displaying the detection result of the contact parameter.

2. The device for detecting parameters of contact points of a proximity detector based on image processing according to claim 1, wherein the processor (7) selects a processing area on the image, filters the interference pixels of the detection area, and obtains an edge image containing the outline of the movable contact point of the proximity detector, specifically as follows:

Converting the region to be processed on the close-fitting inspector image into a gray-scale image in an RGB format, and performing edge detection on the gray-scale image by using a Canny operator to obtain a first edge image containing the contact point of the close-fitting inspector; the area to be treated is an area where the contact group ring exists;

and searching the connected domain at the edge of the first edge image, deleting the connected domain with the number of pixels smaller than a set value, and obtaining a second edge image.

3. The device for detecting the contact point parameter of the close-fitting inspector based on image processing according to claim 1, wherein the processor (7) fits and positions pixel points on an edge image to obtain a movable contact point position image, the device comprises the steps of calibrating a camera by adopting a Zhang Zhengyou plane calibration method, attaching a printed template to a plane in the camera calibration stage, shooting a plurality of template images from different angles, detecting characteristic points in the template images, taking distortion of a camera model into consideration, and solving by utilizing an optimization algorithm to obtain accurate internal and external parameters and distortion coefficients of the camera.

4. The device for detecting parameters of contact points of a contact checker based on image processing according to claim 1, wherein the display (8) is used for displaying the detection results output by the processor (7), including the image of the start and end states of the moving contact, the driving depth of the moving contact ring and the pressure of the static contact at the driving moving contact.

5. The method for detecting the contact point parameters of the close-fitting inspector based on image processing is characterized by comprising the following steps:

Step1, arranging a first camera (1) and a second camera (2) in a box body of the close-fitting inspector, and detecting contact points of the close-fitting inspector by adopting a non-contact method;

Step 2, observing the internal condition of the box body of the close-fitting inspector through the first camera (1) and the second camera (2), shooting the moving state of the movable contact of the close-fitting inspector, and storing the images of the start state and the end state;

Step 3, a processor (7) collects the start and end state images of the close-fitting inspector to be processed, splices the images and selects an area to be processed on the close-fitting inspector images;

step 4, filtering the interference pixel points of the area to be processed to obtain an edge image containing the outline of the movable contact of the close-fitting inspector;

Step 5, fitting, positioning and correcting distortion of pixel points on the edge image to obtain a movable contact point position image;

Step 6, processing the obtained movable contact point position image of the close-contact detector, and positioning the center coordinates and the diameter of a movable contact point moving ring through extracting the characteristics of the movable contact point to obtain the contact point parameters of the close-contact detector, wherein the contact point parameters comprise the contact point group pressure of the close-contact detector, the depth of the movable contact point connecting with the moving ring of the close-contact detector and the displacement distance of the moving ring of the working start and stop of the close-contact detector, and the method comprises the following steps of:

Firstly calculating the gradient of each pixel point in an image, traversing non-0 pixel points in the image after edge detection, drawing a straight line along the gradient of the non-0 pixel points, accumulating points through which the straight line passes, accumulating points with larger points in an accumulator being more likely to be circle centers, reserving points which are larger than a given threshold value and are larger than all neighbors of the points in the accumulator as candidate points, and arranging the candidate points in descending order according to accumulated values so as to ensure that the point most likely to be the circle center appears earliest; after selecting circle center candidate points, considering all non-0 pixels for each center, and making a radius for a possible circle center;

Positioning the edge of a movable joint moving ring of the close-fitting inspector in the image through the relevant parameter setting of Hough circle detection; the circle center position and the diameter of the movable joint movable ring are positioned through the extraction of the circle center radius; acquiring the opening distance of the static contact sheet by detecting the diameter of the movable ring, and carrying in the functional relation between the opening distance and the stress to obtain the contact group pressure of the close-fitting inspector; obtaining the pixel distance of the moving ring by closely attaching the position of the circle center of the moving ring of the inspector to the starting and ending states, and obtaining the actual distance of the moving ring and the driving depth of the moving ring by calibrating a camera; the detection of the close-fitting inspector is converted into digital calibration from manual abstract observation.

6. The method for detecting contact point parameters of a patch inspector based on image processing according to claim 5, wherein the filtering of the interference pixels of the area to be processed in step 4 obtains an edge image including a contour of a movable contact point of the patch inspector, specifically comprising the following steps:

Converting the region to be processed on the close-fitting inspector image into a gray-scale image in an RGB format, and performing edge detection on the gray-scale image by using a Canny operator to obtain a first edge image containing the contact point of the close-fitting inspector; the area to be treated is an area where the contact group ring exists;

and searching the connected domain at the edge of the first edge image, deleting the connected domain with the number of pixels smaller than a set value, and obtaining a second edge image.

7. The method for detecting contact point parameters of a close-fitting inspector based on image processing according to claim 5, wherein in step 5, pixel points on an edge image are fitted, positioned and corrected for distortion to obtain a moving contact point position image, the method comprises the steps of calibrating a camera by adopting a Zhang Zhengyou plane calibration method, attaching a printed template to a plane in a camera calibration stage, shooting a plurality of template images from different angles, detecting characteristic points in the template images, considering distortion of a camera model, and solving by utilizing an optimization algorithm to obtain accurate camera internal and external parameters and distortion coefficients.