A vehicle-mounted highway guardrail beam plate bolt absence rapid detection device

Abstract

The application provides a kind of vehicle-mounted highway guardrail beam plate bolt missing quick detection device, including mounting mechanism, camera system, light supplementing lamp system, control system, wherein mounting mechanism is used to and the connection between vehicle, camera system is connected on mounting mechanism and carries out data transmission between control system, camera system is used to realize the collection of guardrail image, light supplementing lamp system is connected on mounting mechanism, to light for camera system, control system is used to the guardrail image data collected by camera system is analyzed and handled, also be used to the control of camera system and light supplementing lamp system simultaneously.By the device to the bolt missing condition of highway guardrail is detected, can greatly improve the efficiency and speed of detection, avoid closed road traffic, while realizing the comprehensive detection of guardrail, avoid the situation that the error of sampling detection result is larger.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traffic safety facilities detection, and particularly relates to a vehicle-mounted highway guardrail beam plate bolt loss rapid detection device. BACKGROUND

[0002] Wave-shaped beam steel guardrails are very common in highway engineering, which are composed of wave-shaped beam plates, upright columns, end heads, splicing bolts, connecting bolts, blocking blocks, brackets, cross beams and other components. The wave-shaped beam plates and upright columns are two important components of the wave-shaped beam steel guardrails, which have a very important influence on the mechanical properties and normal functioning of the protective function of the guardrails. Since the guardrails are arranged along the longitudinal direction of the highway, the length is often long, so the wave-shaped beam plates need to be connected by splicing bolts to form a complete wave-shaped beam plate system, and the wave-shaped beam plates also need to be connected with the guardrails. The connection between the wave-shaped beam plates and the guardrails mainly depends on the connecting bolts. If the splicing bolts or the connecting bolts are loose or fall off during use, it will seriously affect the connection between the wave-shaped beam plates and the upright columns, thereby affecting the crashworthiness of the entire guardrail and posing a serious threat to traffic safety.

[0003] Therefore, during the operation of the highway, it is necessary to check and count the missing situation of the guardrail bolts in order to supplement and improve them in a timely manner and ensure the mechanical properties of the guardrail. However, since the number of splicing bolts and connecting bolts in the guardrail is large, it is time-consuming and laborious to check them one by one, so during the actual inspection and road maintenance, the detection personnel generally randomly select 5 beam plates with bolts per kilometer for detection, and obtain the missing situation of the selected bolts by visual inspection. The biggest defect of the sampling inspection mode is that the number of selected samples is small and not representative, and selecting more samples will make the detection work time-consuming and laborious and low in efficiency. Moreover, the bolt is relatively small compared to the entire guardrail system, and it is also limited by the special structure of the wave-shaped beam plate, making it difficult to obtain the missing situation of the highway guardrail bolts through other ways.

[0004] Therefore, the current detection of bolt loss of highway wave-shaped beam steel guardrails mainly has the following problems: 1. The manual detection is low in efficiency, long in detection time and small in sampling quantity, and the traffic needs to be closed during detection, thereby affecting the normal traffic of the road; 2. The rapid and accurate detection cannot be realized, especially in the process of detecting the highway, the emergency lane needs to be occupied, thereby threatening the personal safety of the detection personnel; 3. There is no tool and method for rapid detection of bolt loss of guardrails in the prior art, and most of the bolt loss detection means are based on unmanned aerial vehicles and mainly applied in the field of power inspection, which is not applicable in the field of highway engineering. Therefore, it is urgent to propose a device and method capable of efficiently and safely detecting the bolt loss of wave-shaped beam steel guardrails. SUMMARY

[0005] In view of the problems of low efficiency and safety hazards in the traditional manual detection of the bolt loss of highway wave-shaped beam steel guardrails, the application provides a vehicle-mounted rapid detection device for bolt loss of guardrail beam plates, which realizes rapid detection of the bolt loss of guardrails based on a vehicle-mounted device, does not affect the normal traffic of vehicles on the road, and can effectively ensure the detection efficiency and the safety of the detection personnel.

[0006] To solve the above technical problems, the application realizes the following technical scheme: a vehicle-mounted rapid detection device for bolt loss of highway guardrail beam plates, comprising a mounting mechanism, a camera system, a fill light system and a control system, wherein the mounting mechanism is used for connection with a vehicle, the camera system is connected to the mounting mechanism and transmits data with the control system, the camera system is used for image acquisition of guardrails, the fill light system is connected to the mounting mechanism to provide light for the camera system, and the control system is used for analysis and processing of the image data of the guardrails collected by the camera system and for control of the camera system and the fill light system. In actual application, the vehicle-mounted rapid detection device for bolt loss of highway guardrail beam plates is first fixed and installed on one side of the vehicle close to the guardrail through the mounting mechanism, then the vehicle normally travels on the road section to be detected, the control system controls the camera system to start and begin to collect image data of the guardrails, the camera system transmits the collected image data of the guardrails to the control system, the control system controls the start and stop of the fill light system according to the image information of the guardrails, and after the detection of all the guardrails is completed, the control system outputs the bolt loss situation of the wave-shaped beam steel guardrails on the road section.

[0007] Further, the mounting mechanism further comprises a height adjusting mechanism and a mounting platform, one end of the height adjusting mechanism is fixedly connected to the mounting platform, the other end of the height adjusting mechanism is fixedly connected to the vehicle, and the height adjusting mechanism can adjust the mounting height of the mounting platform on the vehicle, so that the vehicle-mounted rapid detection device for bolt loss of highway guardrail beam plates can adapt to vehicles of different heights.

[0008] Further, the camera system comprises a camera angle adjusting mechanism, a high-dynamic camera, the camera angle adjusting mechanism is fixedly installed above the installation platform, and the high-dynamic camera is rotationally connected to the camera angle adjusting mechanism. By setting the camera angle adjusting mechanism, the angle of the camera can be adjusted to adapt to the image acquisition requirements of different road conditions and different beam type guardrails. The sampling frequency of the high-dynamic camera is 25 Hz, and the sampling interval is 2 m. ， The high-dynamic camera has an image collection field of 1.4*0.7 m.

[0009] Further, the light supplement system comprises a light supplement angle adjusting mechanism and a light supplement. The light supplement angle adjusting mechanism is fixedly installed below the installation platform, and the light supplement is rotationally connected to the light supplement angle adjusting mechanism. By setting the light supplement angle adjusting mechanism, the angle of the light supplement can be adjusted to better adapt to the camera, so that the light supplement can better meet the light supplement requirements of the high-dynamic camera.

[0010] Further, the control system comprises an upper computer, a lower computer, a GPS, a relay, and a distance sensor. The upper computer is composed of data acquisition software and data processing software. The high-dynamic camera collects data in real time, and sends images to the upper computer through Ethernet. The GPS sends positioning information to the lower computer through a serial port, and the distance sensor sends distance data to the lower computer through an I / O port. After data shaping by the lower computer, the lower computer sends the positioning information and distance information to the upper computer through Ethernet. The upper computer sends a light-on instruction to the lower computer in combination with the brightness information of the image, and the lower computer controls the opening and closing of the light supplement through the relay.

[0011] The upper computer combines traditional vision and deep learning vision in the visual algorithm. The traditional vision part is responsible for preprocessing the image to remove noise of the high-dynamic camera, water mist in the environment, suppress backlight, and adjust the brightness of the image. The deep learning part is responsible for positioning detection of the guardrail beam and detection of the bolt.

[0012] The guardrail beam detection network and the guardrail post detection network are trained using the shallow network of YOLOV5, and the inference is performed using the 16-bit inference engine of TensorRT. The bolt detection network is trained using the deep network of YOLOV5, and the inference is performed using the 16-bit inference engine of TensorRT. The detected guardrail posts are processed for duplicate detection. The image of the same post is packaged into a data packet and placed in the analysis queue. Each data packet in the queue is detected using the guardrail bolt detection network. The image of the post with the most detected bolts is determined as the bolt detection result of the post, including the bolt, missing hole, and contour mark.

[0013] The detailed steps of the upper computer for detection are as follows:

[0014] Image preprocessing includes noise removal, image length and width adjustment, dark channel dehazing, image brightness adjustment, image normalization, and image channel adjustment.

[0015] Image B is copied from memory to video memory for inference of the guardrail beam detection network model and the column detection network model, and non-maximum suppression processing is applied to the detection results, while the detection results located at the image edge are removed.

[0016] C performs duplicate detection on the pillars as follows: Record the vehicle's mileage D1 for the first detected pillar, located at column coordinate X1 in the image; record the vehicle's mileage D for subsequent detected pillars. M The column coordinate X in the image M The value of M ranges from 2 to 100.

[0017] The pillars on the left side of the car are considered to be the same pillar if they satisfy formulas (C-1) and (C-2); the pillars on the right side of the car are considered to be the same pillar if they satisfy formulas (C-1) and (C-3).

[0018] D M -D1<2(C-1)

[0019] X M -X1≤100(C-2)

[0020] X M -X1≥-100(C-3)

[0021] The images captured from the same column are packaged into a single data packet and placed into the queue to be analyzed.

[0022] D copies the captured images from the data packets in the queue to be analyzed from memory to video memory for bolt detection model inference, performs non-maximum suppression processing on the detection results, and removes the detection results located at the image edges. The captured image of the column with the most detected guardrail bolts is identified as the bolt detection result of that column.

[0023] E obtains the bolt inspection results, which include the inspection data for bolts, missing holes, and delineators;

[0024] F is used to calculate the missing rate, where the missing rate = number of missing holes / (number of bolts + number of missing holes + number of delineators).

[0025] The host computer is an industrial control computer or a personal computer, and the slave computer is a PLC or MCU.

[0026] The beneficial effects of this invention are as follows:

[0027] 1. The device uses a vehicle-mounted device, and the installation mechanism, camera system, light supplement system and control system are arranged on the vehicle, which will not affect the normal traffic of the road during the detection process. The detection process does not require contact with the guardrail, nor does it require manual on-site measurement. The detection speed can reach 90km / h at the highest;

[0028] 2. The device can achieve full detection of the guardrail. The traditional guardrail bolt missing detection method is sampling, which needs to balance the detection efficiency and accuracy. Since the length of the guardrail is often long, but the position of the bolt missing situation is uncertain, it is impossible to achieve full coverage inspection of the guardrail bolt, and some missing bolt guardrails may be missed during the detection process. The device based on machine vision and image recognition can sample and analyze all guardrails on the detected section without consuming much time.

[0029] 3. The device can better adapt to the highway guardrail. The device uses a vehicle as a carrying device and installs the device on the side of the vehicle. The vehicle and the guardrail maintain a certain distance, and the height and angle of the high dynamic camera and the light supplement lamp can be adjusted through the height adjustment mechanism and the angle adjustment mechanism, so that the device as a whole can be applied to various grades of highways. When detecting on the highway, since the emergency lane is set on the roadside, the distance between the device and the guardrail is large, and the adjustable range of the device installation height and the camera installation angle is large, which can better ensure the size of the high dynamic camera image acquisition. When detecting on ordinary national roads, provincial roads, municipal roads or rural roads, since there is generally no emergency lane, the distance between the device and the guardrail is close, and the height of the device is adjusted reasonably to keep it consistent with the height of the guardrail beam plate, and the angle of the high dynamic camera and the light supplement lamp is adjusted reasonably to effectively ensure the size and quality of the high dynamic camera image acquisition. Compared with the image acquisition method using a drone, due to the special structure of the wave-shaped guardrail, the bolt is generally installed in the concave part of the guardrail beam plate. The drone shooting is affected by trees, vehicles and the guardrail itself, which is not conducive to the acquisition of the guardrail image. The device is closer to the guardrail and has better continuity in image acquisition, and the acquired image is more convenient for analysis and processing.

[0030] 3. Reduce the number of detection personnel. The traditional detection method requires the arrangement of multiple on-site detection personnel, while the device only needs one detection personnel and one driver to carry out detection work. The detection personnel can complete the on-site detection by operating the supporting software on the vehicle, and the detection personnel can detect on the vehicle, effectively avoiding the risk factors of on-site detection and better protecting the personal safety of the detection personnel;

[0031] 4. Higher detection accuracy, the traditional manual sampling method can only realize the estimation of the missing situation of the guardrail bolt in the sampling section, the device can more accurately obtain the detection result of the missing situation of the guardrail bolt through comprehensive sampling of the guardrail and the de-duplication algorithm, and the accuracy of the bolt missing detection of the device can reach 95% through verification.

[0032] 5. Easy to review, since the device collects images of all guardrails of the road section to be tested, the abnormal detection result in the detection process can be reviewed again through manual review of the corresponding image. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is the overall structure diagram of the vehicle-mounted highway guardrail beam plate bolt missing rapid detection device of the present application;

[0034] Figure 2 is the front view of the vehicle-mounted highway guardrail beam plate bolt missing rapid detection device of the present application;

[0035] Figure 3 is the side view of the vehicle-mounted highway guardrail beam plate bolt missing rapid detection device of the present application;

[0036] Figure 4 is the visual algorithm flowchart of the present application;

[0037] Figure 5 is the system structure diagram of the present application;

[0038] Figure 6 is the overall detection result output by the host computer of the present application;

[0039] Figure 7 is the detection result counted by the host computer according to the post number of the present application;

[0040] Figure 8 is the guardrail image identified by the present application;

[0041] Figure 9 is the bolt identification situation in the collection process of the present application.

[0042] Wherein: 1- mounting mechanism; 2- camera system; 3- fill light system; 4- height adjusting mechanism; 5- mounting platform; 6- camera angle adjusting mechanism; 7- high dynamic camera; 8- fill light angle adjusting mechanism; 9- fill light. DETAILED DESCRIPTION

[0043] In order to deepen the understanding of the present application, the present application will be further described in combination with the embodiments below, and the present embodiment is only used to explain the present application and does not constitute a limitation on the protection scope of the present application.

[0044] EMBODIMENT

[0045] As Figures 1-6 shown, the embodiment proposes a vehicle-mounted highway guardrail beam plate bolt missing rapid detection device, including mounting mechanism 1, camera system 2, light supplement system 3, control system, wherein the mounting mechanism is used for the connection between the vehicle, the camera system is connected to the mounting mechanism and transmits data between the control system, the camera system is used to realize the collection of guardrail image, the light supplement system is connected to the mounting mechanism, which is used to supplement light for the camera system, the control system is used for analyzing and processing the guardrail image data collected by the camera system, and is also used for controlling the camera system and the light supplement system. In the actual application process, first of all, through the mounting mechanism, the vehicle-mounted highway guardrail beam plate bolt missing rapid detection device is fixedly installed on the side of the vehicle close to the guardrail, then the vehicle normally drives on the road section to be detected, the control system controls the camera system to start and begin to collect guardrail image data, the camera system transmits the collected guardrail image data to the control system, the control system controls the start and stop of the light supplement system according to the guardrail image information, and after the detection of all guardrails is completed, the control system outputs the missing situation of the bolt of the wave-shaped beam steel guardrail of the detected road section.

[0046] Further, the mounting mechanism further includes a height adjusting mechanism 4 and a mounting platform 5, one end of the height adjusting mechanism is fixedly connected to the mounting platform, the other end of the height adjusting mechanism is fixedly connected to the vehicle, and the height adjusting mechanism can adjust the mounting height of the mounting platform on the vehicle, so that the vehicle-mounted highway guardrail beam plate bolt missing rapid detection device can adapt to vehicles of different heights. The height adjusting mechanism can adopt an electric cylinder, a telescopic rod, etc.

[0047] Further, the camera system includes a camera angle adjusting mechanism 6 and a high dynamic camera 7, the camera angle adjusting mechanism is fixedly installed above the mounting platform, and the high dynamic camera is rotatably connected to the camera angle adjusting mechanism. By setting the camera angle adjusting mechanism, the angle of the camera can be adjusted, so that the camera can adapt to the image collection requirements of different road conditions and different beam plate type guardrails. The camera angle adjusting mechanism 6 includes a U-shaped block, two rotating plates, two rotating rods, two guide rods and a nut, two vertical plates are symmetrically arranged on the two sides of the U-shaped block, an arc-shaped guide groove and a circular rotating groove are formed in the vertical plates, the two rotating plates are fixedly connected to the two sides of the high dynamic camera 7, so that the high dynamic camera 7 is clamped between the two vertical plates, one end of each of the two rotating rods is fixedly connected to the corresponding rotating plate, the other end of each of the two rotating rods is inserted into the circular rotating groove of the vertical plate and rotatably connected to the vertical plate, one end of each of the two guide rods is fixedly connected to the corresponding rotating plate, the other end of each of the two guide rods is inserted into the arc-shaped guide groove of the vertical plate and slidably connected to the arc-shaped guide groove, the guide rod adopts a screw rod, and after the angle of the high dynamic camera 7 is adjusted in place, the position of the guide rod is fixed by using the nut, thereby fixing the position and angle of the high dynamic camera 7.

[0048] Further, the light supplement system comprises a light supplement angle adjusting mechanism 8 and a light supplement 9, the light supplement angle adjusting mechanism is fixedly installed below the installation platform, and the light supplement is rotationally connected to the light supplement angle adjusting mechanism. By setting the light supplement angle adjusting mechanism, the angle of the light supplement can be adjusted to better adapt to the camera, so that the light supplement can better meet the light supplement demand of the high dynamic camera. The sampling frequency of the high dynamic camera is 25Hz, and the sampling interval is 2m ， The image size collected by the high dynamic camera is 1.4*0.7m. According to the guardrail design specification, the interval between adjacent posts of the corrugated beam guardrail is 1m, 1.5m, 2m, 3m or 4m. If the camera sampling interval is 1m or 1.5m, the sampling interval is too short, which will cause the collected guardrail image data to be too large, and the image processing efficiency will be reduced. If the camera sampling interval is set to 3m or 4m, the sampling interval is too long, which will cause part of the guardrail image to be missed, and full coverage detection cannot be achieved. Therefore, the sampling interval of the high dynamic camera is determined as the median value of the post interval, i.e. 2m, which can ensure full coverage of image collection and appropriately reduce the data amount of the image, facilitating analysis and processing.

[0049] Further, the control system comprises an upper computer, a lower computer, a GPS, a relay and a distance sensor. The upper computer is composed of data acquisition software and data processing software. The high dynamic camera collects real-time data and sends images to the upper computer through Ethernet. The GPS sends positioning information to the lower computer through a serial port, and the distance sensor sends distance data to the lower computer through an I / O port. After data shaping by the lower computer, the lower computer sends the positioning information and distance information to the upper computer through Ethernet. The upper computer sends the light-on command to the lower computer in combination with the brightness information of the image, and the lower computer controls the opening and closing of the light supplement through the relay.

[0050] Among them, the upper computer adopts an industrial computer or a personal computer, and the lower computer adopts a PLC or an MCU.

[0051] The upper computer combines traditional vision and deep learning vision in the visual algorithm. The traditional vision part is responsible for preprocessing the image, removing noise of the high dynamic camera, water mist in the environment, suppressing backlight, and adjusting the brightness of the image. The deep learning part is responsible for positioning detection of the guardrail beam and detection of the bolt.

[0052] The guardrail beam plate detection network and the guardrail post detection network are trained using a shallow network of YOLOV5, and are inferred using a 16-bit inference engine of TensorRT. The bolt detection network is trained using a deep network of YOLOV5, and is inferred using a 16-bit inference engine of TensorRT. The detected guardrail posts are subjected to a duplicate detection process, the images of the same post are packaged into a data packet and put into the analysis queue, each data packet in the queue is detected using the guardrail bolt detection network, and the image of the post with the most detected bolts, missing holes and contour marks is determined as the bolt detection result of the post.

[0053] The detailed steps of the host computer for detection are as follows:

[0054] A. Image preprocessing, i.e. removing noise, adjusting image length and width, image dark channel defogging, adjusting image brightness, image normalization, and image channel adjustment;

[0055] B. Copy the image from the memory to the video memory for guardrail beam plate detection network model and post detection network model inference, and perform non-maximum suppression on the detection results, while removing the detection results located at the edge of the image;

[0056] C. De-duplicate the detected posts, the specific operation is as follows: record the vehicle mileage D1 of the first detected post, the column coordinate X1 in the image, record the vehicle mileage D M of the subsequent detected post, the column coordinate X M in the image, and M is in the range of 2 to 100.

[0057] When the left post of the car satisfies formula (C-1) and formula (C-2), it is determined as the same post; when the right post of the car satisfies formula (C-1) and formula (C-3), it is determined as the same post.

[0058] D M -D1<2(C-1)

[0059] X M -X1≤100(C-2)

[0060] X M -X1≥-100(C-3)

[0061] Pack the images of the same post into a data packet and put it into the analysis queue;

[0062] For the above column de-weighting step, the device triggers a high dynamic camera to sample at a distance of 2 m, and uses formula (C-1) to calculate the difference in vehicle mileage to determine whether the collected images are within the same sampling interval range. If the detected column vehicle mileage difference is less than 2 m, it indicates that the collected images are within the same sampling interval range, and the column in the image may have a repetition. Further, according to the specification, the minimum spacing of the guardrail column is 1 m, and formula (C-2) is used to calculate the spacing between the columns in the collected images. The spacing is calculated using the difference in column coordinates in the image, and if the absolute value of the difference in column coordinates of the two columns in the image is less than 100 cm (i.e., 1 m), it indicates that there are repeated columns in the two images.

[0063] D Copy the intercepted image in the data packet in the queue to be analyzed from the memory to the video memory for bolt detection model inference, and perform non-maximum suppression processing on the detection results, while eliminating the detection results located at the edge of the image, and determine the column intercepted image with the most detected guardrail bolt results as the bolt detection result of the column;

[0064] E Get the bolt detection result, which includes bolt, missing hole, and contour mark detection data;

[0065] F Calculate the missing rate, where the missing rate = missing hole number / (bolt number + missing hole number + contour mark number).

[0066] Through field testing and verification, when the vehicle speed is not more than 90 km / h, the accuracy of the bolt missing detection of the device can reach more than 95%, and when the speed exceeds 90 km / h, the accuracy of the bolt missing detection will decrease. When the vehicle speed is 90 km / h, the corresponding speed converted to m / s is 25 m / s, and according to the specification, the minimum spacing of the guardrail steel column is 1 m. At this time, the sampling frequency of the high dynamic camera is determined to be 25 Hz, which can ensure that at least one image is collected at an average distance of 1 m. This effectively avoids the large amount of data collected by high-frequency sampling of the camera while ensuring the number of guardrail image collections, ensuring that the device can process data in real time, balancing precision and efficiency.

[0067] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and the scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A vehicle-mounted device for rapid detection of bolt absence of highway guardrail beam plates, characterized in that: The device comprises a mounting mechanism, a camera system, a light supplement system, and a control system, wherein the mounting mechanism is used for connection with a vehicle, the camera system is connected to the mounting mechanism and performs data transmission with the control system, the camera system is used for image acquisition of a guardrail, the light supplement system is connected to the mounting mechanism and provides light supplement for the camera system, and the control system is used for analysis and processing of image data of the guardrail acquired by the camera system and control of the camera system and the light supplement system; The mounting mechanism is used for fixing and mounting the device on one side of the vehicle close to the guardrail, the vehicle drives on a road section to be detected, the control system controls the camera system to start and begin image acquisition of the guardrail, the camera system transmits the acquired image data of the guardrail to the control system, the control system controls the light supplement system to start and stop according to the image information of the guardrail, and after detection of all the guardrails is completed, the control system outputs the missing situation of the bolts of the wave-shaped beam steel guardrail of the detected road section; The mounting mechanism further comprises a height adjusting mechanism and a mounting platform, one end of the height adjusting mechanism is fixedly connected to the mounting platform, and the other end of the height adjusting mechanism is fixedly connected to the vehicle; The control system comprises an upper computer, a lower computer, a GPS, a relay, and a distance sensor, the upper computer comprises data acquisition software and data processing software, a high dynamic camera acquires real-time data, and the image is sent to the upper computer through Ethernet; the GPS sends positioning information to the lower computer through a serial port, the distance sensor sends distance data to the lower computer through an I / O port, the positioning information and the distance information are sent to the upper computer through Ethernet after being shaped by the lower computer, the upper computer sends a light-on instruction to the lower computer in combination with the brightness information of the image, and the lower computer controls the light supplement system to open and close through the relay; The upper computer combines a traditional visual algorithm and a deep learning visual algorithm in terms of visual algorithm, wherein the traditional visual algorithm is responsible for image preprocessing, removal of noise of the high dynamic camera, water mist in the environment, suppression of backlight, and adjustment of brightness of the image; and the deep learning visual algorithm is responsible for detection of a guardrail beam plate, detection of a guardrail post, and detection of a bolt; The guardrail beam plate detection network and the guardrail post detection network are trained using a shallow network of YOLOV5 and are inferred using a 16-bit inference engine of TensorRT, and the bolt detection network is trained using a deep network of YOLOV5 and is inferred using a 16-bit inference engine of TensorRT; The detailed steps of detection realized by the upper computer are as follows: A. image preprocessing, including removal of noise, adjustment of length and width of the image, image dark channel defogging, adjustment of brightness of the image, image normalization, and image channel adjustment; B. copying the image from the memory to the video memory for inference of a guardrail beam plate detection network model and a post detection network model, performing non-maximum suppression processing on the detection results, and eliminating the detection results located at the edge of the image; C. The detected posts are de-duplicated, as follows: the car mileage D1 at which the first detected post is recorded, the column coordinate X1 in the image, the car mileage D at which the subsequent detected posts are recorded, and the column coordinate X in the image are recorded M , M is in the range of 2 to 100 M ; When the left post of the automobile satisfies formula (C-1) and formula (C-2), it is determined as the same post; and when the right post of the automobile satisfies formula (C-1) and formula (C-3), it is determined as the same post. D M -D1<2 (C-1) X M - X1≤ 100 (C-2) X M - X1≥ -100 (C-3) Packaging the intercepted image of the same column into a data packet and putting it into the analysis queue; D Copying the intercepted image in the data packet in the analysis queue from the memory to the video memory for bolt detection model reasoning, performing non-maximum suppression processing on the detection result, eliminating the detection result located at the edge of the image, and identifying the column intercepted image with the most detected guardrail bolts as the bolt detection result of the column; E Obtaining the bolt detection result, which includes the detection data of the bolt, the missing hole, and the contour mark; F Calculating the missing rate, wherein the missing rate = missing hole number / (bolt number + missing hole number + contour mark number).

2. A vehicle mounted quick bolt absence detection device for highway guardrail beam plates according to claim 1, characterized in that, The camera system comprises a camera angle adjusting mechanism and a high-dynamic camera, the camera angle adjusting mechanism is fixedly installed above the installation platform, and the high-dynamic camera is rotationally connected to the camera angle adjusting mechanism, the sampling frequency of the high-dynamic camera is 25 Hz, the sampling interval is 2 m, and the image field of view of the high-dynamic camera is 1.4*0.7 m.

3. A vehicle mounted highway guardrail beam bolt absence quick detection device according to claim 1, characterized in that, The light supplement lamp system comprises a light supplement lamp angle adjusting mechanism and a light supplement lamp, the light supplement lamp angle adjusting mechanism is fixedly installed below the installation platform, and the light supplement lamp is rotationally connected to the light supplement lamp angle adjusting mechanism, and the number of the light supplement lamp is one or two.

4. A vehicle mounted quick bolt absence detection device for highway guardrail beam plates according to claim 1, characterized in that, The upper computer is an industrial computer or a personal computer, and the lower computer is a PLC or an MCU.

Images