Implementation method of AI leaving alarm system for flat shunting service

Abstract

The invention discloses an implementation method of an AI leaving alarm system for plane shunting service. The system comprises an AI processing module, an image acquisition module and a display terminal module. The system collects information of a shunting worker portrait recognition picture frame in an image collection area in real time and sends the information to the AI processing module through a local area network, and the AI processing module compares the proportion of the portrait recognition picture frame fixed by a shunting worker in the collection area with the proportion of the portrait recognition picture frame fixed by the shunting worker in a flat shunting driver seat. And calculating and judging the proportion of the portrait recognition picture frame which is acquired in real time in the moving direction of the shunting personnel in the image acquisition area, further judging whether to give an alarm or not according to the proportion, and sending alarm information to a background locomotive leading personnel display terminal for real-time display. By the adoption of the system, operation accidents possibly caused by the fact that a shunting worker leaves a driver seat and cannot receive the plane shunting service sent by an operation background train leading worker in the railway plane shunting service can be eradicated or avoided, and therefore driving safety of the plane shunting service is guaranteed.

Description

Technical Field

[0001] This invention relates to railway horizontal shunting systems, and more particularly to a method for implementing an AI-based off-duty alarm system for horizontal shunting operations. Background Technology

[0002] Railway shunting operations require shunting operators to remain in the driver's seat. If the shunting operator leaves the driver's seat and fails to receive the shunting instructions from the control center in a timely manner, it may lead to an operational accident. Summary of the Invention

[0003] In surface shunting operations, if the shunting operator leaves the driver's seat for any reason, the vehicle will be left unattended, which can easily lead to accidents. The purpose of this invention is to provide real-time status information of the shunting operator in the driver's seat for surface shunting operations. When the shunting operator leaves the driver's seat, an alarm signal is sent to the display terminal of the back-end dispatcher, and the shunting operator's driver's seat flashes an alarm light and emits an alarm sound.

[0004] The technical solution adopted by this invention is: a method for implementing an AI-based off-duty alarm system for flatbed shunting operations. The AI-based off-duty alarm system for flatbed shunting operations consists of an image acquisition module, an AI processing module, and a display terminal. First, the system acquires images of shunting operators in the driver's seat of the flatbed shunting operation through the image acquisition module within the image acquisition area, obtaining information about the fixed facial recognition frame of the shunting operator in the driver's seat within the image acquisition area. Then, it acquires information about the facial recognition frame of the shunting operator moving left and right within the image acquisition area in real time, and sends the acquired information about the shunting operator's facial recognition frame within the image acquisition area to the AI ​​processing module through a local area network. The AI ​​processing module compares the proportion occupied by the fixed facial recognition frame of the shunting operator in the driver's seat within the acquisition area, calculates and determines the proportion of the facial recognition frame after the real-time acquisition of the shunting operator's movement direction within the image acquisition area, and then determines whether to issue an alarm based on the proportion, and sends the alarm information to the back-end dispatcher's display terminal for real-time display.

[0005] The AI ​​processing module of this invention generates real-time images from the real-time collected facial recognition frame information of the shunting operator. If the occupancy ratio of the moved shunting operator's facial recognition frame is 50%, it is determined that the shunting operator is in place, and no alarm message is sent. If the occupancy ratio of the shunting operator's facial recognition frame is less than 50% but greater than 30%, it is determined that the shunting operator's operation is not standardized, but he is still in place. Then, an alarm message is sent to the back-end leader's display terminal, and the shunting operator's driver's seat alarm light flashes to remind the shunting operator to pay attention to receiving crane services from the leader, but no alarm sound is played. If the occupancy ratio of the shunting operator's facial recognition frame is less than 30%, it is determined that the shunting operator has left the driver's seat. Then, an alarm message is sent to the back-end leader's display terminal, the leader stops sending crane services, the shunting operator's driver's seat alarm light flashes and an alarm sound is played to remind the shunting operator to return to the driver's seat.

[0006] The beneficial effects of this invention are: the use of this system can eliminate or avoid operational accidents that may occur in railway shunting operations due to the shunting operator leaving the driver's seat and failing to receive the shunting orders sent by the lead operator in a timely manner, thereby ensuring the driving safety of shunting operations. Attached Figure Description

[0007] Figure 1 This is a structural block diagram of the AI-based off-duty alarm system for flatbed shunting operations according to the present invention; Figure 2 This is a flowchart illustrating the principle of the AI-based off-duty alarm system for flatbed shunting operations according to the present invention. Figure 3 This is a schematic diagram of the AI ​​processing module algorithm analyzing images according to the present invention. Detailed Implementation

[0008] The present invention will be further described below with reference to the accompanying drawings.

[0009] like Figure 1 , Figure 2 As shown, the AI-powered off-duty alarm system for horizontal shunting operations consists of an image acquisition module, an AI processing module, and a display terminal. The image acquisition module collects video information from the driver's seat during horizontal shunting, the AI ​​processing module analyzes the video images from the driver's seat, and the display terminal module displays real-time alarm information, thereby ensuring driving safety in horizontal shunting operations.

[0010] First, the system acquires images of the shunting operator in the driver's seat of the flat shunting yard through the image acquisition module within the image acquisition area. This obtains information about the fixed facial recognition frame of the shunting operator in the image acquisition area. Then, it acquires information about the facial recognition frame of the shunting operator moving left and right within the image acquisition area in real time. This information is then sent to the AI ​​processing module via the local area network. The AI ​​processing module compares the proportion of the fixed facial recognition frame of the shunting operator in the acquisition area with the proportion of the fixed facial recognition frame in the driver's seat. It calculates and determines the proportion of the facial recognition frame after the shunting operator's movement direction is captured in real time. Based on this proportion, it determines whether to issue an alarm and sends the alarm information to the back-end dispatcher's display terminal for real-time display.

[0011] The AI ​​processing module of this invention generates real-time images from the real-time collected facial recognition frame information of the shunting operator. If the occupancy ratio of the moved shunting operator's facial recognition frame is 50%, it is determined that the shunting operator is in place, and no alarm message is sent. If the occupancy ratio of the shunting operator's facial recognition frame is less than 50% but greater than 30%, it is determined that the shunting operator's operation is not standardized, but he is still in place. Then, an alarm message is sent to the back-end leader's display terminal, and the shunting operator's driver's seat alarm light flashes to remind the shunting operator to pay attention to receiving crane services from the leader, but no alarm sound is played. If the occupancy ratio of the shunting operator's facial recognition frame is less than 30%, it is determined that the shunting operator has left the driver's seat. Then, an alarm message is sent to the back-end leader's display terminal, the leader stops sending crane services, the shunting operator's driver's seat alarm light flashes and an alarm sound is played to remind the shunting operator to return to the driver's seat.

[0012] The image acquisition module is used to collect video information from the driver's seat during shunting operations, the AI ​​processing module is used to analyze the video images from the driver's seat, and the display terminal module is used to display real-time alarm information, thereby ensuring driving safety in shunting operations.

[0013] The image acquisition module of the AI-powered off-duty alarm system for flat-plane shunting operations consists of a network infrared camera, the AI ​​processing module consists of an Atlas 200I AI acceleration module and core algorithms, and the display terminal consists of a computer and display terminal software displayed on it.

[0014] like Figure 3 As shown, the image acquisition area is 1920x1080 resolution. Figure 3 The shunting operator's facial recognition frame in the middle has a resolution of 1280x900 and occupies approximately 55% of the image. Therefore, it should be determined that the shunting operator is present, and no alarm message should be sent. However... Figure 3If the proportion of the shunting operator's facial recognition frame, which indicates that he is moving to the left, is less than 50% but greater than 30%, it should be judged that the shunting operator's operation is not standardized, but he is still in position. Therefore, an alarm message is sent to the back-end dispatcher's display terminal, and the shunting operator's driver's seat alarm light flashes to remind the shunting operator to pay attention to receiving crane services from the dispatcher, but no alarm sound is played. Figure 3 If the proportion of the shunting operator's facial recognition frame moving to the right is less than 30%, it should be determined that the shunting operator has left the driver's seat. An alarm message should then be sent to the back-end dispatcher's display terminal. The dispatcher should stop sending crane services, and the alarm light on the shunting operator's driver's seat should flash and play an alarm sound to remind the shunting operator to return to the driver's seat.

[0015] The high-definition infrared camera in the image acquisition module captures real-time video information from the driver's seat and sends it to the AI ​​processing module via a local area network. The AI ​​processing module generates real-time images based on the video information. AI algorithms determine whether the shunting operator has left the driver's seat and their presence status by judging the proportion of the captured image area occupied by the operator's image. When the shunting operator moves within the driver's seat, the vertical movement of the image is negligible; the primary comparison is made regarding horizontal movement.

Claims

1. A method for implementing an AI-based off-duty alarm system for flatbed shunting operations, characterized in that, The AI-powered off-duty alarm system for shunting operations consists of an image acquisition module, an AI processing module, and a display terminal. First, the system acquires images of the shunting operator in the driver's seat within the image acquisition area using the image acquisition module. This obtains information about the fixed facial recognition frame of the shunting operator within the image acquisition area. Then, it acquires information about the shunting operator's left-right moving facial recognition frame within the image acquisition area in real time. This acquired information is then sent to the AI ​​processing module via a local area network. The AI ​​processing module compares this information with the proportion occupied by the fixed facial recognition frame of the shunting operator in the driver's seat within the acquisition area. It calculates and determines the proportion of the facial recognition frame after the operator's movement direction is detected, and then determines whether to issue an alarm based on this proportion. The alarm information is then sent to the back-end dispatcher's display terminal for real-time display.

2. The implementation method of an AI-based off-duty alarm system for flatbed shunting operations according to claim 1, characterized in that, The AI ​​processing module generates real-time images from the shunting operator's facial recognition frame information collected in real time. If the moved shunting operator's facial recognition frame occupies 50% of the frame, the shunting operator is considered to be in place, and no alarm message is sent. If the shunting operator's facial recognition frame occupies less than 50% but more than 30%, the shunting operator's operation is considered to be non-standard, but the operator is still in place. An alarm message is then sent to the backend dispatcher's display terminal, and the shunting operator's driver's seat alarm light flashes to alert the shunting operator to receive crane services from the dispatcher, but no alarm sound is played. If the shunting operator's facial recognition frame occupies less than 30%, the shunting operator has left the driver's seat. An alarm message is then sent to the backend dispatcher's display terminal, the dispatcher stops sending crane services, the shunting operator's driver's seat alarm light flashes, and an alarm sound is played to remind the shunting operator to return to the driver's seat.

Images