A Two-Wheeled Vehicle Elevator Access Prohibition System and Method Based on Deep Learning

Abstract

The present invention discloses a two-wheeled vehicle elevator access prohibition system and method based on deep learning. The video image acquisition unit 2 is used to monitor the interior of the elevator in real time, capture and extract the video images of moving objects. The image preprocessing unit enhances, grayscales, and geometrically transforms the collected video images to ensure the quality of the video. The background change judgment unit judges the background and foreground of the video images. The two-wheeled vehicle detection, recognition, and judgment unit conducts the recognition and judgment of the vehicle. The system feedback braking unit 3 controls the emergency stop braking control of the elevator. The alarm unit 4 gives an alarm reminder when a prohibited access target is detected. The notification unit notifies the staff to coordinate and handle when a prohibited access target is detected. The present invention utilizes the video monitoring system in the elevator to monitor the target objects in the elevator in real time. After detecting and recognizing the target object, the present invention reminds the passengers to abide by the relevant rules through the alarm device of the elevator and timely drives the prohibited access target out of the elevator.

Description

Technical Field

[0001] The present invention relates to the field of anti-misjudgment detection of pedestrians, and particularly to a two-wheeled vehicle elevator access prohibition system and method based on deep learning. Background Art

[0002] With the congestion of roads and the soaring oil prices, electric vehicles and bicycles have become the new favorites for short-distance travel. However, this has brought more potential safety hazards. More and more people drive electric vehicles and bicycles into the elevator and take them to the upper floors, causing potential safety hazards to the elevator and the corridor. The present invention proposes a two-wheeled vehicle elevator access prohibition system and method based on deep learning. When this system detects a prohibited target, it will promptly press the emergency stop button of the elevator to stop the elevator and notify the relevant staff by uploading data to the cloud.

[0003] Currently, the video monitoring system of the elevator only functions to transmit real-time images and save data. Although there are duty personnel watching the video images all the time, due to the excessive number of monitoring screen images and human negligence, many accidents occur without being noticed by the duty personnel in a timely manner. In order to prevent the entry of two-wheeled electric vehicles into the elevator from bringing potential safety hazards to the elevator and passengers at the source, it is very necessary to design an elevator target access prohibition system. Summary of the Invention

[0004] The purpose of the present invention is to provide a two-wheeled vehicle elevator access prohibition system and method based on deep learning, to solve the problem that electric vehicles and bicycles enter the elevator and the corridor due to negligence in personnel management, thus causing the occurrence of electric vehicle spontaneous combustion and corridor blockage, and to achieve the safety of taking the elevator.

[0005] The technical solution adopted by the present invention is:

[0006] A two-wheel vehicle elevator access control system based on deep learning, which includes a control center and a video image acquisition unit, a system feedback braking unit, and an alarm unit connected to the control center; the control center includes a main control board, and an image preprocessing unit, a background change judgment unit, a two-wheel vehicle detection and recognition judgment unit, and a notification unit are carried on the main control board; the control center is connected to the video image acquisition unit, the system feedback braking unit, and the alarm unit through the main control board; wherein, the video image acquisition unit monitors the elevator interior in real time, captures and extracts the video images of moving objects; the image preprocessing unit enhances, grayscales, and geometrically transforms the acquired video images to ensure the quality of the video; the background change judgment unit uses the inter-frame difference method to judge the background and foreground of the video image to determine whether there is a vehicle target entering the elevator; the two-wheel vehicle detection and recognition judgment unit performs vehicle recognition and judgment based on a neural network; the system feedback braking unit controls the emergency stop braking of the elevator; the alarm unit gives an alarm reminder when a prohibited access target is detected; the notification unit notifies the staff to coordinate and handle when a prohibited access target is detected.

[0007] Further, after the video images collected by the video image acquisition unit are converted into digital signals through A / D conversion, they are compressed and encoded using Divx to generate an MPEG4 data stream.

[0008] Further, the inter-frame difference method judges whether it is in motion or the background by calculating the difference in the gray values of two adjacent frames of images, that is, judges the background and foreground of the video image. The inter-frame difference method has strong adaptability to light, and the algorithm is simple and the processing speed is fast.

[0009] Further, the neural network is the YOLO V4 detection network.

[0010] Specifically, the input end of the YOLO V4 detection network uses the Mosaic method for data augmentation to enrich the data and increase the robustness of the network; the Backbone of the YOLO V4 detection network is changed from CSPDarknet53 to MobileNet V3 to make the network lightweight and more in line with the detection and recognition requirements; two new PAN structures are added to the Neck structure of the YOLO V4 detection network. The SPP unit can make any size of input image be unified into a fixed size when output, increasing the receptive field. Through the SPP+PANet structure for upsampling and downsampling, the predicted feature maps can be better fused, further improving the network's feature extraction ability; the spatial-wise attention mechanism of the SAM unit in the YOLO V4 detection network becomes a point-wise attention mechanism, and the shortcut connection in the PAN unit becomes a concatenation connection.

[0011] Furthermore, the system feedback braking unit includes an external switch, which is electrically connected to the control center. The external switch has an external armature, and the external armature of the external switch corresponds to the built-in armature of the emergency stop switch of the elevator. When the external switch is turned on, the external armature is energized and attracts the built-in armature, so that a pair of normally open contacts of the emergency stop switch are closed to realize the emergency stop of the elevator.

[0012] Furthermore, the alarm unit includes a PLC main control board and a buzzer, and the PLC main control board transmits a current signal to the buzzer so that the buzzer emits a sound alarm to remind passengers.

[0013] Furthermore, the notification unit uses the TCP / IP protocol to upload and synchronize the alarm data to the cloud server so as to remind the staff through the cloud server.

[0014] A method for prohibiting two-wheeled vehicles from entering elevators based on deep learning, the specific steps are as follows:

[0015] Step 1, collecting real-time image data of the elevator interior and the elevator entrance through an image acquisition unit;

[0016] Step 2, preprocessing the extracted video image through an image preprocessing unit to improve the quality and confidence of the video image;

[0017] Step 3, the background change judgment unit judges whether there is a background change; if yes, extract the video image with background change and pass it to the detection and recognition unit and execute step 4; otherwise, execute step 1;

[0018] Step 4: The detection and recognition unit uses the trained and mature neural network to determine whether there is a prohibited target object; if yes, execute step 5; otherwise, execute step 1;

[0019] Step 5: The control center sends an electrical signal to control the alarm and notification unit to send out an alarm to remind passengers to evacuate; at the same time, the control center sends an electrical signal to the control system feedback brake unit to brake the elevator, so that the elevator stops running immediately;

[0020] Step 6: The control center uploads the alarm data to the cloud through the notification unit to notify the staff to go to the site in time for coordination and processing.

[0021] Furthermore, the banned objects include bicycles and electric vehicles.

[0022] Furthermore, the alarm buzzer selects the Qiante LTD-1101 model magnetic rotating warning light, which can work under 12V AC power, and the magnetic base can be adsorbed on the steel surface for easy installation.

[0023] Further, upload the prohibited target data to the PC side and synchronize it to the cloud. Compress and encode the video or picture data captured and confirmed as the prohibited target to generate a data stream.

[0024] Further, in step 6, when the prohibited target exits the elevator within the set time limit, the notification unit does not upload the alarm data to the cloud and does not notify the staff to go to the scene in time for coordination and handling; when the prohibited target does not exit the elevator within the set time limit, the notification unit uploads the alarm data to the cloud and notifies the staff to go to the scene in time for coordination and handling.

[0025] Further, the set time limit is 10 seconds.

[0026] The present invention adopts the above technical solutions to use the video monitoring system in the elevator to monitor the elevator target objects in real time; when the detected and recognized target is an electric motorcycle, a bicycle, or an electric bicycle, the system will control the elevator to stop running immediately through a feedback signal; after detecting and recognizing the target object, the elevator's alarm device is used to remind passengers to abide by relevant rules and promptly drive the prohibited target out of the elevator; for the situation where the prohibited target has not been driven out of the elevator within 10 seconds, the system will send the relevant situation to the relevant staff to go to the scene in time for coordination and handling. Description of the Drawings

[0027] The following further elaborates on the present invention in detail in conjunction with the drawings and specific embodiments;

[0028] Figure 1 It is a flowchart of a method for prohibiting two-wheeled vehicles from entering an elevator based on deep learning according to the present invention;

[0029] Figure 2 It is a schematic diagram of the overall structure of a system for prohibiting two-wheeled vehicles from entering an elevator based on deep learning according to the present invention;

[0030] Figure 3 It is a schematic block diagram of the control center of a system for prohibiting two-wheeled vehicles from entering an elevator based on deep learning according to the present invention;

[0031] Figure 4 It is a schematic diagram of the principle of the system feedback braking unit of a system for prohibiting two-wheeled vehicles from entering an elevator based on deep learning according to the present invention;

[0032] Figure 5 It is a structural diagram of the system feedback braking unit of a system for prohibiting two-wheeled vehicles from entering an elevator based on deep learning according to the present invention. Detailed Embodiments

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0034] AsFigures 1 to 5 As shown in one of them, the present invention discloses a two-wheel vehicle elevator access prohibition system based on deep learning, which includes a control center and a video image acquisition unit 2, a system feedback braking unit 3, and an alarm unit 4 connected to the control center 1; the control center 1 includes a main control board 5, and an image preprocessing unit, a background change judgment unit, a two-wheel vehicle detection and recognition judgment unit, and a notification unit are mounted on the main control board 5; the control center 1 is connected to the video image acquisition unit 2, the system feedback braking unit 3, and the alarm unit 4 through the main control board 5; wherein, the video image acquisition unit 2 monitors the elevator interior in real time, captures and extracts the video images of moving objects; the image preprocessing unit enhances, grayscales, and geometrically transforms the acquired video images to ensure the video quality; the background change judgment unit uses the inter-frame difference method to judge the background and foreground of the video images to determine whether there is a vehicle target entering the elevator; the two-wheel vehicle detection and recognition judgment unit identifies and judges the vehicle based on a neural network; the system feedback braking unit 3 controls the emergency stop braking control of the elevator; the alarm unit 4 gives an alarm reminder when a prohibited access target is detected; the notification unit notifies the staff to coordinate and handle when a prohibited access target is detected.

[0035] Further, the video images collected by the video image acquisition unit 2 are converted into digital signals through A / D conversion, compressed, and encoded using Divx to generate an MPEG4 data stream.

[0036] Further, the inter-frame difference method judges whether it is in motion or the background by calculating the difference in grayscale values between two adjacent frames of images, that is, judges the background and foreground of the video images. The inter-frame difference method has strong adaptability to light, and the algorithm is simple and the processing speed is fast.

[0037] Further, the neural network is a YOLO V4 detection network.

[0038] Specifically, the input end of the YOLO V4 detection network uses the Mosaic method for data augmentation to enrich the data and increase the network robustness; the Backbone of the YOLO V4 detection network is changed from CSPDarknet53 to MobileNet V3 to make the network lightweight and more in line with the detection and recognition requirements; two new PAN structures are added to the Neck structure of the YOLO V4 detection network. The SPP unit can unify any size of input image into a fixed size when outputting, increase the receptive field, and perform upsampling and downsampling through the SPP+PANet structure to better fuse the prediction feature maps, further improving the network's feature extraction ability; the spatial-wise attention mechanism of the SAM unit in the YOLO V4 detection network becomes a point-wise attention mechanism, and the shortcut connection in the PAN unit becomes a concatenation connection.

[0039] Furthermore, the system feedback brake unit 3 includes an external switch 6, which is electrically connected to the main control board 5 of the control center 1. The external switch 6 has an external armature, and the external armature of the external switch 6 corresponds to the built-in armature of the emergency stop switch 7 of the elevator. When the external switch is turned on, the external armature is energized and attracts the built-in armature to turn on the emergency stop switch 7, so that a pair of normally open contacts 9 of the emergency stop switch 7 are closed to realize the emergency stop of the elevator. As a feasible implementation, the external switch 6 can be set inside or outside the elevator; it can also be set in the monitoring room of the control center.

[0040] The emergency stop switch 7 is an internal structure of the elevator, and the present invention does not modify it. The electrical signal triggers the armature normally open contact 9 to close, thereby achieving the purpose of emergency stop of the elevator. The external switch 6 controls the emergency stop button in the same principle. The main control board 5 and the external switch 6 transmit signals through a wired connection.

[0041] System feedback braking unit 3 braking process principle: When the control center 1 detects and identifies the forbidden target, the control center 1 sends an electrical signal through the main control board 5 to control the armature of the external switch 6 and the armature of the emergency stop button to attract each other, so that a pair of normally open contacts 9 are closed, so as to achieve the purpose of pressing the elevator emergency stop button through the external switch 6. After the normally open contact 9 between the external switch 6 and the emergency stop button is closed, the normally open contact 9 inside the emergency stop button will be closed, thereby controlling the elevator to stop running.

[0042] Furthermore, the alarm unit 4 includes a PLC main control board and a buzzer. The PLC main control board transmits a current signal to the buzzer so that the buzzer emits a sound alarm to remind the passengers.

[0043] Furthermore, the notification unit uses the TCP / IP protocol to upload and synchronize the alarm data to the cloud server 8 so as to remind the staff through the cloud server 8.

[0044] A method for prohibiting two-wheeled vehicles from entering elevators based on deep learning, the specific steps are as follows:

[0045] Step 1, collecting real-time image data of the elevator interior and the elevator entrance through an image acquisition unit;

[0046] Step 2, preprocessing the extracted video image through an image preprocessing unit to improve the quality and confidence of the video image;

[0047] Step 3, the background change judgment unit judges whether there is a background change; if yes, extract the video image with background change and pass it to the detection and recognition unit and execute step 4; otherwise, execute step 1;

[0048] Step 4, the detection and recognition unit uses the trained and mature neural network to determine whether there is a prohibited entry target; if so, execute Step 5; otherwise, execute Step 1;

[0049] Step 5, the control center 1 sends an electrical signal to control the alarm and notification unit to issue an alarm to remind passengers to evacuate; at the same time, the control center 1 sends an electrical signal to the system feedback braking unit 3 to brake the elevator, so that the elevator stops running immediately;

[0050] Step 6, the control center 1 uploads the alarm data to the cloud through the notification unit to notify the staff to arrive at the scene in time for coordination and handling.

[0051] Further, the prohibited entry targets include bicycles and electric vehicles.

[0052] Further, the alarm buzzer selects the magnetic rotation warning light of the Kent LTD-1101 model, which can work under 12V alternating current, and the magnetic base can be adsorbed on the steel surface for easy installation.

[0053] Further, upload the prohibited entry target data to the PC side and synchronize it to the cloud, and compress and encode the video or picture data captured and confirmed as the prohibited entry target to generate a data stream.

[0054] Further, in Step 6, when the prohibited entry target drives out of the elevator within the set time limit, the notification unit does not upload the alarm data to the cloud and does not notify the staff to arrive at the scene in time for coordination and handling; when the prohibited entry target does not drive out of the elevator within the set time limit, the notification unit uploads the alarm data to the cloud and notifies the staff to arrive at the scene in time for coordination and handling.

[0055] Further, the set time limit is 10 seconds.

[0056] The following is a detailed description of the specific principle of the present invention:

[0057] The present invention mainly forms an elevator prohibited entry system through the layout of each unit device, constructs a method for detecting, recognizing, alarming and braking electric vehicles and bicycles entering the elevator. For the conventional technical means not within the scope of the technical solution of the present invention, it is not necessary to refine the equipment model, installation and configuration method, each unit component, etc. of each device in this embodiment.

[0058] As Figure 3 shown, the elevator prohibited entry system performs overall data analysis and processing by the control center 1; it is composed of a video image acquisition unit 2, an image preprocessing unit, a background change judgment unit, a detection and recognition unit, an alarm unit, a feedback braking unit and a notification unit. The device is normally powered on, and the system starts to monitor the moving objects inside the elevator and at the elevator entrance in real time and capture the moving object image data.

[0059] Example 1

[0060] The present invention intends to implement the monitoring and recognition results when only passengers and no prohibited entry targets enter the elevator.

[0061] First, the image acquisition unit collects real-time image data inside the elevator and at the elevator entrance. The image preprocessing unit preprocesses the extracted video images to improve the quality and confidence of the video images.

[0062] The background change judgment unit judges the background, determines that the background has indeed changed, and transmits the extracted video images to the detection and recognition unit.

[0063] The detection and recognition unit determines that no prohibited entry target is detected.

[0064] In Example 1, there are only passengers, and no electric vehicles and bicycles are detected as prohibited entry targets. The elevator access control system then returns to continue collecting and extracting video images.

[0065] Example 2

[0066] The present invention intends to implement the monitoring and judgment results when a passenger drives an electric vehicle into the elevator.

[0067] First, the image acquisition unit collects real-time image data inside the elevator and at the elevator entrance. The image preprocessing unit preprocesses the extracted video images to improve the quality and confidence of the video images.

[0068] The background change judgment unit judges the background, determines that the background has indeed changed, and transmits the extracted video images to the detection and recognition unit.

[0069] The detection and recognition unit determines that an electric vehicle is detected.

[0070] The control center 1 sends an electrical signal to control the alarm unit to issue an alarm to remind passengers to evacuate.

[0071] The control center 1 sends an electrical signal to control the feedback braking unit 07 to immediately press the peripheral switch 6 to stop the elevator immediately.

[0072] The control center 1 sends an electrical signal to control the unit 08 for notifying relevant staff to notify the staff to arrive at the scene in time for coordination and handling by uploading data to the cloud.

[0073] In Example 2, a passenger drives an electric vehicle into the elevator, and it is detected and recognized that an electric vehicle has entered the elevator. The elevator access control system alarms, stops urgently, and notifies the staff.

[0074] Example 3

[0075] The present invention intends to implement the monitoring and judgment results when a passenger drives a bicycle into the elevator.

[0076] First, the image acquisition unit collects real-time image data inside the elevator and at the elevator entrance. The image preprocessing unit preprocesses the extracted video images to improve the quality and confidence of the video images.

[0077] The background change judgment unit judges the background, determines that the background has indeed changed, and transmits the extracted video images to the detection and recognition unit.

[0078] The detection and recognition unit determines that a bicycle has been detected.

[0079] The control center 1 sends an electrical signal to control the alarm unit to issue an alarm to remind passengers to evacuate.

[0080] The control center 1 sends an electrical signal to control the feedback braking unit 07 to immediately press the peripheral switch 6 to stop the elevator immediately.

[0081] The control center 1 sends an electrical signal to control the unit 08 for notifying relevant staff to upload data to the cloud to notify the staff to arrive at the scene in time for coordination and handling.

[0082] In the second embodiment, a passenger drives a bicycle into the elevator, and it is detected and recognized that a bicycle has entered the elevator. The elevator access prohibition system alarms, stops urgently, and notifies the staff.

[0083] The present invention adopts the above technical solutions to use the video monitoring system in the elevator to monitor the elevator target objects in real time; when it is detected and recognized that the target is an electric motorcycle, a bicycle, or an electric bicycle, the system will control the elevator to stop immediately through a feedback signal; after detecting and recognizing the target object, the alarm device of the elevator will remind passengers to comply with relevant regulations and promptly drive the prohibited target out of the elevator; for the situation where the prohibited target has not been driven out of the elevator within 10 seconds, the system will send relevant situations to relevant staff to arrive at the scene in time for coordination and handling.

[0084] Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Without conflict, the embodiments and features in the embodiments of the present application can be combined with each other. Usually, the components of the embodiments of the present application described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the present application claimed, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

Claims

1. A two-wheel vehicle elevator access prohibition system based on deep learning, characterized in that: It includes a control center and a video image acquisition unit, a system feedback braking unit, and an alarm unit connected to the control center; the control center includes a main control board, on which an image preprocessing unit, a background change judgment unit, a two-wheel vehicle detection and recognition judgment unit, and a notification unit are mounted; the control center is connected to the video image acquisition unit, the system feedback braking unit, and the alarm unit through the main control board; among them, the video image acquisition unit monitors the elevator interior in real time, captures and extracts the video images of moving objects; the image preprocessing unit enhances, grayscales, and geometrically transforms the collected video images to ensure the quality of the video; the background change judgment unit uses the inter-frame difference method to judge the background and foreground of the video image to determine whether there is a vehicle target entering the elevator; the two-wheel vehicle detection and recognition judgment unit performs vehicle recognition and judgment based on a neural network; the system feedback braking unit controls the emergency stop braking of the elevator; the alarm unit gives an alarm reminder when a prohibited access target is detected; the notification unit notifies the staff to coordinate and handle when a prohibited access target is detected; among them, the neural network is the YOLO V4 detection network; the input end of the YOLO V4 detection network uses the Mosaic method for data augmentation to enrich the data and increase the robustness of the network; the Backbone of the YOLO V4 detection network is changed from CSPDarknet53 to MobileNet V3 to make the network lightweight and more in line with the detection and recognition requirements; two new PAN structures are added to the Neck structure of the YOLO V4 detection network, and the SPP unit is used to make any size of the input image unified into a fixed size when output to increase the receptive field, and up and down sampling are performed through the SPP+PANet structure to better fuse the predicted feature maps and further improve the feature extraction ability of the network; the spatial-wise attention mechanism of the SAM unit in the YOLO V4 detection network becomes a point-wise attention mechanism, and the shortcut connection in the PAN unit becomes a concatenation connection.

2. A two-wheel vehicle elevator access prohibition system based on deep learning according to claim 1, characterized in that: After the video image collected by the video image acquisition unit is converted into a digital signal through A / D, it is compressed and encoded using Divx to generate an MPEG4 data stream.

3. A two-wheel vehicle elevator access prohibition system based on deep learning according to claim 1, characterized in that: The inter-frame difference method judges whether it is in motion or the background by calculating the difference in the gray values of two adjacent frames of images, that is, judges the background and foreground of the video image. The inter-frame difference method has strong adaptability to light, and the algorithm is simple and the processing speed is fast.

4. A two-wheel vehicle elevator access prohibition system based on deep learning according to claim 1, characterized in that: The system feedback braking unit includes an external switch, which is electrically connected to the control center. The external switch has an external armature, and the external armature of the external switch corresponds to the built-in armature of the emergency stop switch of the elevator. When the external switch is turned on, the external armature is energized and attracts the built-in armature, so that a pair of normally open contacts of the emergency stop switch are closed to realize the emergency stop of the elevator.

5. According to the deep learning-based two-wheeled vehicle elevator entry prohibition system of claim 1, Features: The alarm unit includes a PLC main control board and a buzzer. The PLC main control board transmits a current signal to the buzzer so that the buzzer emits a sound alarm to remind passengers.

6. A two-wheeled vehicle elevator entry prohibition system based on deep learning according to claim 1, Features: The notification unit uses TCP / IP protocol to upload and synchronize alarm data to the cloud server so that the staff can be reminded through the cloud server.

7. A two-wheeled vehicle elevator entry prohibition method based on deep learning, using a two-wheeled vehicle elevator entry prohibition system based on deep learning according to any one of claims 1 to 6, Features: The method comprises the following steps: Step 1, collecting real-time image data of the elevator interior and the elevator entrance through an image acquisition unit; Step 2, preprocessing the extracted video image through an image preprocessing unit to improve the quality and confidence of the video image; Step 3, the background change judgment unit judges whether there is a background change; if yes, extracts the video image with background changes and transmits it to the detection and recognition unit and executes step 4; Otherwise, go to step 1; Step 4: The detection and recognition unit uses the trained and mature neural network to determine whether there is a prohibited target object; if yes, execute step 5; otherwise, execute step 1; Step 5: The control center sends an electrical signal to control the alarm and notification unit to send out an alarm to remind passengers to evacuate; at the same time, the control center sends an electrical signal to the control system feedback brake unit to brake the elevator, so that the elevator stops running immediately; Step 6: The control center uploads the alarm data to the cloud through the notification unit to notify the staff to go to the site in time for coordination and processing.

8. A two-wheeled vehicle elevator entry prohibition method based on deep learning according to claim 7, Features: Step 6: When the prohibited target leaves the elevator within the set time limit, the notification unit does not upload the alarm data to the cloud, and does not notify the staff to go to the scene in time for coordination and processing; when the prohibited target does not leave the elevator within the set time limit, the notification unit uploads the alarm data to the cloud, and notifies the staff to go to the scene in time for coordination and processing.

9. A two-wheeled vehicle elevator entry prohibition method based on deep learning according to claim 8, Features: Set the time limit to 10 seconds.

Images