Vehicle-mounted full-scene video intelligent acquisition and line detection device

Abstract

The utility model discloses a vehicle-mounted full-scene video intelligent acquisition and line detection device, which solves the problem of lack of safety monitoring and detection equipment for railway locomotives running along the line. The device comprises a video acquisition unit, an acceleration acquisition unit, a coordinate acquisition unit, a storage unit, an interface unit, a man-machine interaction unit and a main control circuit board, the video acquisition unit comprises a front camera, a left camera and a right camera which are respectively mounted in front of and on two sides of a cab of a locomotive, and the front camera is connected with the left camera. Video acquisition is carried out on monitoring areas on the two sides of the railway track; the acceleration acquisition unit comprises a three-axis accelerometer, is mounted on a locomotive and is used for acquiring data of a low-frequency shaking vehicle; the acceleration sensor of the device adopts a triaxial accelerometer with low noise, low null drift and low power consumption, has a good low-frequency characteristic, is more sensitive to low-frequency vehicle shaking caused by line irregularity, and greatly improves the accuracy and consistency of line detection.

Description

Technical Field

[0001] The utility model relates to the technical field of railway line safety operation and monitoring. Background Art

[0002] Railway line equipment is the basic equipment of the railway transportation industry. Because it is exposed to nature all year round, it is subjected to the effects of wind, rain, freeze-thaw and the impact of train operation, which causes the line quality status to change constantly. The roadbed and ballast are constantly deformed, and the rails, connecting parts and sleepers are constantly worn, causing the state of the line equipment to change constantly. The lack of railway protection nets leads to the invasion of foreign objects, and the power supply cables and signal equipment around the line are damaged by natural factors, resulting in abnormal operation, which in turn affects driving safety. Therefore, real-time monitoring of the line environment and surrounding supporting equipment is conducive to timely and accurate detection of the location of the disease, and then notifying relevant departments for timely maintenance and replacement, which can reasonably avoid driving accidents and provide strong guarantees for safe railway operations.

[0003] Currently, there is no intelligent means for real-time monitoring and disease detection along the railway lines where locomotives operate. Summary of the invention

[0004] In order to solve the problems described in the background technology, the utility model provides a vehicle-mounted full-scene video intelligent acquisition and line detection device to solve the problem of lack of safety monitoring and detection equipment along the railway locomotive operation line.

[0005] The technical solution adopted by the utility model to solve its technical problems is:

[0006] The vehicle-mounted full-scene video intelligent acquisition and line detection device includes a video acquisition unit, an acceleration acquisition unit, a coordinate acquisition unit, a storage unit, an interface unit, a human-computer interaction unit and a main control circuit board, and is characterized in that:

[0007] The video acquisition unit includes a front camera, a left camera and a right camera, which are respectively installed in front of and on both sides of the locomotive cockpit to collect video of the monitoring areas on both sides of the railway track;

[0008] The acceleration acquisition unit includes a three-axis accelerometer and is installed on the locomotive to collect data on low-frequency vehicle shaking;

[0009] The coordinate acquisition unit includes a Beidou positioning module and is installed on the locomotive to locate the locomotive;

[0010] The storage unit is composed of a TF card and an EMMC memory;

[0011] The interface unit is composed of a data transmission interface, a power supply interface and an antenna interface;

[0012] The transmission unit is located on the main control circuit board, and the transmission unit is composed of a Bluetooth module, a WLAN module and a 4G / 5G full network module;

[0013] The human-computer interaction unit displays the line curve, slope, vibration acceleration curve and working interface;

[0014] The video acquisition unit, the acceleration acquisition unit, the coordinate acquisition unit, the storage unit, the interface unit, and the human-computer interaction unit are electrically connected to the main control circuit board respectively.

[0015] Furthermore, the main control circuit board is an RK3588 core board, and a data acquisition module and an AI analysis module are integrated on the main control circuit board.

[0016] Furthermore, the AI ​​analysis module is composed of an AI detection submodule and an AI analysis storage submodule.

[0017] Furthermore, the human-computer interaction unit is a mobile phone or a tablet.

[0018] Furthermore, the human-machine interaction unit is wirelessly connected to the main control circuit board via wireless communication.

[0019] Furthermore, the video acquisition unit adopts a 2560*1440P / 60FPS high-definition camera.

[0020] The beneficial effects of the utility model are:

[0021] The acceleration sensor of this device adopts a low-noise, low-zero drift, and low-power three-axis accelerometer, which has good low-frequency characteristics and is more sensitive to low-frequency shaking caused by uneven lines, greatly improving the accuracy and consistency of line detection; using the Beidou positioning module, fast positioning, stable satellite search, can achieve real-time, accurate positioning, speed measurement, automatic identification of line name, line type and the acquisition of information of the management boundary. Share data with the locomotive through Bluetooth communication to accurately identify the location of the disease; have 2K or 4K full-scene high-definition video recording, storage, operation information superposition display function, convenient for later over-limit data analysis, and accurately locate the mileage of the disease, and can also make manual marks for other abnormal behaviors of the line, which is convenient for on-site handling of related problems; have touch display function, clear display, flexible operation; have line curve, slope, and vibration acceleration curve display function, real-time display of curve, slope and vibration waveform information at the running mileage; share the added information through the APP in the mobile terminal, you can view the added information in real time on the mobile phone, and you can also remotely view the equipment over-limit and added line track information; use a low-power portable printer to quickly print current and historical over-limit information.

[0022] This device can not only overcome the shortcomings and defects of traditional line dynamic detection, but also combine new technologies to reasonably apply data acquisition, image acquisition and AI artificial intelligence analysis technology in the field of railway line dynamic detection technology. It can not only detect line diseases, but also monitor the supporting equipment around the line (power supply equipment, signal equipment) in real time throughout the whole process. It uses WLAN and Bluetooth communication technology to realize remote return of key data, and combines AI artificial intelligence analysis technology to timely and effectively analyze the working conditions of line supporting equipment. It can meet the use of various types of locomotives. The equipment is simple to operate and the handheld terminal is easy to carry, which has certain advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a system design block diagram of the utility model.

[0024] Figure 2 Design a block diagram for the transmission unit. DETAILED DESCRIPTION

[0025] The on-board full-scene video intelligent acquisition and line detection device includes multiple unit modules such as the main control circuit board, video acquisition unit, acceleration acquisition unit, coordinate acquisition unit, storage unit, interface unit, human-computer interaction unit, etc., which can be configured to monitor the scenes along the railway in real time.

[0026] This device adopts modular design concept in hardware design, and its system design block diagram is as follows: Figure 1 shown.

[0027] The video acquisition unit, acceleration acquisition unit and coordinate acquisition unit are respectively used to acquire video data, acceleration data and coordinate data during the locomotive's forward movement, and are also called acquisition units.

[0028] The video acquisition unit includes three cameras, namely the front camera, the left camera and the right camera, which are installed in front of and on both sides of the locomotive cockpit. They are used to collect video of the monitoring area within N meters on both sides of the railway track. They can record and monitor in real time, obtain basic video data, and perform AI analysis on the technical data to find abnormal parts of the scene. The acquisition unit in the device is equipped with a high-definition camera that can record the environmental information in front of and on both sides of the locomotive in real time to ensure the clarity and stability of the video image. At the same time, appropriate lenses and angles are used to ensure that the video can cover each important monitoring point.

[0029] The above-mentioned high-definition camera is fixedly installed as follows: bracket + power module + camera. The camera is placed on the top of the bracket and its direction is adjustable.

[0030] Furthermore, the above-mentioned high-definition camera adopts a 2560*1440P / 60FPS high-definition camera. This type of camera can not only realize high-definition video acquisition of the environmental information in front of the locomotive, but also can simultaneously save the running line name, line type, mileage information and video information. That is, it has the function of automatic superposition of running information, which can superimpose the running information (running line name, line type, mileage information, etc.) into the video, so that the video has data blessing; the storage time of a single video file is 5 minutes, and it will be cyclically overwritten after exceeding the preset space; the over-limit data is associated with the video information, and clicking on the over-limit data will jump to the video information of the corresponding mileage, which can intuitively display the environmental information of the over-limit alarm location and realize the accurate positioning of the disease location; it has the function of emergency video stream interception or screenshot, so that the video evidence can be fixed, which is convenient for timely and accurate elimination of positioning line diseases and line environment abnormalities.

[0031] Acceleration acquisition unit: It uses a low-noise, low-zero-drift, low-power three-axis accelerometer installed on the locomotive. It has good low-frequency characteristics and is more sensitive to low-frequency vehicle shaking caused by uneven lines. It can more truly reflect the actual conditions of the line, greatly improving the accuracy and consistency of line detection;

[0032] Coordinate acquisition unit: Using Beidou high-precision positioning module or mobile phone / tablet (built-in Beidou high-precision positioning module) to assist in positioning, it can quickly and accurately obtain the current position of the locomotive; 1. Using Beidou positioning mode, the satellite search is stable, and it can achieve real-time and accurate positioning, speed measurement, automatic identification of line name, line type and its management boundary information; 2. Using mobile phone / tablet to assist in positioning, assist the positioning module to quickly locate, and accurately correct the mileage in real time.

[0033] The storage unit is composed of a TF card and an EMMC memory, which is used to store real-time detection data and corresponding video data. The TF card uses a waterproof, heat-resistant, and impact-resistant SanDisk high-speed memory card with a reading speed of 190MB / s and a writing speed of 130MB, which can quickly read and write detection data. The EMMC memory is built into the RK3588 core board with a capacity of 8GB, which is used for key data backup storage to prevent the loss of key test data due to special reasons causing TF card abnormalities.

[0034] The interface unit consists of a data transmission interface, a power supply interface and an antenna interface, and is electrically connected to the RK3588 core board. The data transmission interface is used for communication between the host and the host computer, which can receive parameter information set by the host computer and export data stored in the storage unit to the host computer. The power supply interface is used to connect to the locomotive power supply interface, and the internal integrated power management module provides working power for the host. The antenna interface is used to connect the multi-frequency positioning antenna corresponding to the Beidou positioning module, so that the host can obtain positioning data in real time.

[0035] The main control circuit board (CPU) adopts the RK3588 core board, and the data acquisition module and AI analysis module are integrated on the main control circuit board.

[0036] The data acquisition module is electrically connected to the video acquisition unit, the acceleration acquisition unit and the coordinate acquisition unit, and is used to acquire panoramic video, locomotive acceleration and locomotive position coordinates.

[0037] The AI ​​analysis module is composed of an AI detection submodule and an AI analysis storage submodule. By performing AI analysis on the video image information collected by the camera frame by frame, the information with abnormal conditions is saved as video and picture files. The line name, line type, mileage and other information are superimposed in the video and picture information, which is conducive to timely and accurate positioning of the actual line after the disease is discovered, and the line disease can be quickly repaired. The AI ​​analysis unit is connected to the acquisition module, and the panoramic video is screened according to the set conditions to obtain video data for locating line diseases and abnormal line environment.

[0038] Furthermore, the implementation process of AI analysis technology is to clean and annotate the collected video data to provide sufficient data sets for AI model training. Use the deep learning framework to build and train the AI ​​model. Improve its recognition accuracy and generalization ability by repeatedly training and optimizing the model. Deploy the trained model to the edge device to achieve real-time monitoring and early warning functions. Through the API interface, integrate the AI ​​model into the existing video surveillance system to achieve automatic analysis and alarm of real-time video.

[0039] Furthermore, in the AI ​​analysis storage module, deep learning technology is used to train an AI model that can identify specific railway events. The model can identify abnormal situations during train travel, such as the presence of idle people, livestock, broken fences on both sides of the track, bird nests on the line, etc. The video images of the camera module are obtained in real time, and these video image information and the positioning information of the detected target are recorded together. Combined with the pre-made target location information (kilometer mark, GPS coordinates, etc.), the AI ​​analysis module is used to screen, compare, analyze, detect and identify the abnormal type of the image, and upload the abnormal information image and alarm information to the background server management system. The real-time video is monitored by the AI ​​model, and once an abnormal situation is found, the early warning system is immediately triggered. The early warning information will be sent to relevant personnel and the results will be saved so that timely measures can be taken to deal with it. The AI ​​model is used to retrospectively analyze historical videos to find potential safety hazards or signs of accidents. This will help railway management departments to prevent accidents and take improvement measures.

[0040] The transmission unit is located on the RK3588 core board and is used to send the panoramic video that meets the set conditions after being screened by the AI ​​analysis unit to the TF card for storage;

[0041] The transmission unit is composed of a Bluetooth module, a WLAN module and a 4G / 5G full network access module. By adopting wireless communication, it can not only realize data sharing with the vehicle-mounted line inspection instrument, accurately identify the location of the disease and solve the positioning problem in the tunnel, but also receive and forward the alarm information of the vehicle-mounted equipment, and upload the added information to the mobile phone / tablet (handheld terminal) in real time. The mobile phone can not only parse and display the data, but also implement wireless control of the portable device. At the same time, the host can connect to the router (or 4G / 5G hotspot) through the WLAN module or remotely transmit the over-limit data and video information to the server through the full network access module, so as to realize remote monitoring of the over-limit information and operation trajectory. In fact, the functional design of the transmission unit is as follows: Figure 2 As shown:

[0042] The collected video images are transmitted to the central server through the 4G / 5G module to ensure transmission speed and data security. At the same time, H264 / H265 video compression technology is used to reduce the size of video data and improve transmission efficiency.

[0043] Furthermore, the locomotive operation data is shared through Bluetooth communication, and the name and mileage of the line on which the locomotive is running are automatically identified by combining the route recognition algorithm and the mileage correction algorithm, thereby reducing the location of the defect and improving maintenance efficiency. It can not only share the host detection data, but also remotely control the host working mode, with flexible operation and easy use. It can be used for manual marking and coordinate collection functions, and can remotely control and record different marking information and the current actual line mileage position in real time.

[0044] By connecting the host to a router or mobile phone hotspot through WLAN communication, real-time monitoring data can be transmitted to a remote terminal to achieve remote data sharing; by wirelessly connecting the host to the camera, you can watch the current video information in real time and access historical videos, while avoiding the difficulties of wired wiring and cumbersome operations.

[0045] Through 4G / 5G communication, data sharing between the host and the remote terminal is realized, and video information and detection data are uploaded to the server through the 4G / 5G network. The line environment and mileage location of the defect can be viewed in real time, which is conducive to timely and effective emergency handling of line abnormalities; construction plans and reminder information can be issued remotely to assist additional crew members to complete the construction plan and improve work efficiency.

[0046] An AI analysis unit has been added to screen the acquired panoramic videos and send videos with abnormal vehicles to the TF card for storage, avoiding the TF card storing too many useless videos and affecting the memory space.

[0047] When the locomotive remains stationary, the three-axis accelerometer detects that the vehicle is not vibrating, and the video acquisition module does not record video. When the locomotive starts, the three-axis accelerometer detects and enables the video acquisition unit to collect video data. When harmful video information is detected, the video will be automatically intercepted and sent to the TF card.

[0048] The human-computer interaction unit uses a handheld terminal, such as a tablet or a mobile phone. The handheld terminal mainly uses single-point touch and LCD display technology to display the line curve, slope, vibration acceleration curve and work-related interfaces in real time. At the same time, the operation interface can be manually touched and clicked, with clear display and flexible operation.

[0049] The human-machine interaction unit is wirelessly connected to the main control circuit via wireless communication.

[0050] The design of the on-board panoramic video intelligent acquisition and line detection device combines line dynamic detection and artificial intelligence analysis technology, and realizes the function of viewing real-time data and remote sharing on mobile phones (tablets) through WLAN and Bluetooth wireless communications. The vertical acceleration and horizontal acceleration of the locomotive body vibration are used to judge the line quality disease level and disease type. The line disease type and disease level can be detected timely and accurately. The mileage and video information can be combined to accurately judge the disease location and surrounding environment information, and the disease information can be printed in real time. It has a manual marking function, and the line environment can be manually marked through a remote control or a mobile phone terminal, which is conducive to the timely maintenance of line equipment. The equipment is equipped with high-definition cameras installed in the front and left and right sides of the locomotive, which can record the front and side environmental information in real time. The line name and mileage information recognized by the equipment are transmitted to the camera end through WLAN communication, which is used for AI to accurately obtain the location of the disease when analyzing line abnormalities.

[0051] The embodiments described above are merely descriptions of preferred implementations of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements to the present invention by relevant technical personnel in the field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. A vehicle-mounted full-scene video intelligent acquisition and line detection device, comprising a video acquisition unit, an acceleration acquisition unit, a coordinate acquisition unit, a storage unit, an interface unit, a human-computer interaction unit and a main control circuit board, characterized in that: The video acquisition unit includes a front camera, a left camera and a right camera, which are respectively installed in front of and on both sides of the locomotive cockpit to collect video of the monitoring areas on both sides of the railway track; The acceleration acquisition unit includes a three-axis accelerometer and is installed on the locomotive to collect data on low-frequency vehicle shaking; The coordinate acquisition unit includes a Beidou positioning module and is installed on the locomotive to locate the locomotive; The storage unit is composed of a TF card and an EMMC memory; The interface unit is composed of a data transmission interface, a power supply interface and an antenna interface; The transmission unit is located on the main control circuit board, and the transmission unit is composed of a Bluetooth module, a WLAN module and a 4G / 5G full network module; The human-computer interaction unit displays the line curve, slope, vibration acceleration curve and working interface; The video acquisition unit, the acceleration acquisition unit, the coordinate acquisition unit, the storage unit, the interface unit, and the human-computer interaction unit are electrically connected to the main control circuit board respectively.

2. The vehicle-mounted full-scene video intelligent acquisition and line detection device according to claim 1 is characterized in that: The main control circuit board is an RK3588 core board, and a data acquisition module and an AI analysis module are integrated on the main control circuit board.

3. The vehicle-mounted full-scene video intelligent acquisition and line detection device according to claim 2 is characterized in that: The AI ​​analysis module is composed of an AI detection submodule and an AI analysis storage submodule.

4. The vehicle-mounted full-scene video intelligent acquisition and line detection device according to claim 1 is characterized in that: The human-computer interaction unit is a mobile phone or a tablet.

5. The vehicle-mounted full-scene video intelligent acquisition and line detection device according to claim 1 is characterized in that: The human-machine interaction unit is wirelessly connected to the main control circuit board through wireless communication.

6. The vehicle-mounted full-scene video intelligent acquisition and line detection device according to claim 1 is characterized in that: The video acquisition unit is a high-definition camera.

Images