Subway entrance and exit-oriented state monitoring and alarming system and method

Abstract

The invention relates to a basement entrance and exit-oriented state monitoring and alarming system and method, and the system comprises a preprocessing module which collects an underground garage scene, and a training vehicle detection module; the data acquisition module is used for performing data pulling on the video streams of the monitoring cameras; the edge calculation unit is used for carrying out vehicle detection through a vehicle detection module, matching a detection result of a current frame with a track of a previous frame, and distributing a unique tracking ID; the vehicle data storage module is used for storing the unique tracking ID, the position and the staying time of the vehicle through a hash table structure; the abnormity judgment module is used for counting the vehicle density, the vehicle speed, the relative delay, the current vehicle number and the lane occupation number according to the Hash table structure data, and judging the static, moving, illegal parking and congestion states of the vehicles; and the abnormal state output module is used for sending the real-time information and the log information when the abnormity occurs. Compared with the prior art, the method has the advantages of low cost, easy deployment, high sensitivity and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of basement entrance and exit monitoring, and in particular to a status monitoring and alarm system and method for basement entrances and exits. Background Art

[0002] With the continuous increase in the number of automobiles, underground parking lots have become the main places for vehicle parking. As the only passage for vehicles to enter and exit the underground parking lot, the traffic fluency of the basement entrance and exit has an important impact on the traffic conditions of the entire parking lot and the surrounding roads. Once abnormal events such as traffic congestion or illegal parking occur at the entrance and exit, it will not only lead to a decrease in the operating efficiency of the parking lot, but also may have a serious impact on the surrounding road traffic, further reducing the overall efficiency of regional traffic management.

[0003] Therefore, monitoring events at the basement entrance and exit and providing digital data support for dredging are the keys to ensuring the efficient operation of the parking lot and maintaining the traffic order of the surrounding roads. Existing monitoring methods mainly include manual inspections, retrofitted sensors, and image analysis based on cameras. Among them, manual inspection is a relatively traditional and direct method. The inspection personnel directly observe the queuing situation and driving speed of vehicles at the entrance and exit on the spot, and intuitively judge whether there is congestion, so as to carry out corresponding dredging measures. Retrofitted sensors, such as the technical solution with the publication number CN108288391A, monitor the traffic flow by installing multiple infrared pair sensors on the exit trunk road and using the infrared induction principle. In addition, calculate congestion indicators through existing cameras, such as the technical solution with the authorization announcement number CN114038211B. By monitoring the camera images, calculate the average vehicle delay indicator, which is used as the key basis for judging congestion at the parking lot entrance and exit. This method can achieve automated monitoring to a certain extent, obtain congestion data in a timely manner, and provide strong evidence for subsequent congestion judgment and dredging.

[0004] In summary, the existing basement entrance and exit congestion monitoring technologies have limitations in many aspects. First of all, the manual inspection method is inefficient and has obvious hysteresis, making it difficult to respond to congestion in a timely manner and take effective dredging measures. Secondly, the method of retrofitting sensors or other congestion detection components requires the transformation of existing basement entrances and exits, involving work such as line laying, equipment fixing and debugging. The installation and later maintenance costs are high, making it difficult to widely promote and apply. In addition, a single congestion indicator is difficult to comprehensively and deeply reflect its congestion status. Summary of the Invention

[0005] The purpose of the present invention is to provide a status monitoring and alarm system and method for basement entrances and exits to overcome the defects of the existing technologies, such as the low efficiency and obvious hysteresis of the manual inspection method, and the need to transform existing basement entrances and exits for the method of retrofitting sensors or other congestion detection components.

[0006] The object of the present invention can be achieved by the following technical solutions:

[0007] A state monitoring and alarm system for the entrance and exit of an underground garage, comprising:

[0008] A preprocessing module, configured to collect the underground garage scene, train the vehicle detection module, and calibrate the key monitoring areas of the entrance and exit of the underground garage;

[0009] A data acquisition module, configured to access multiple monitoring cameras and perform parallel data pulling on the video streams of each monitoring camera;

[0010] An edge computing unit, configured to perform vehicle detection through the trained vehicle detection module, match the detection results of the current frame with the previous frame trajectory through the Hungarian matching algorithm, and assign a unique tracking ID;

[0011] A vehicle data storage module, configured to calculate and store the unique tracking ID, position, and residence time of the vehicle through a hash table structure;

[0012] An abnormal situation judgment module, configured to count the vehicle density, vehicle speed, relative delay, current vehicle quantity, and lane occupancy according to the data of the stored hash table structure; when the moving range of the vehicle recognition frame obtained by vehicle detection is within the corresponding moving threshold, it is determined that the vehicle is in a stationary state; if the duration of the vehicle in the stationary state is greater than the corresponding residence threshold, it is determined that the current vehicle is illegally parked; when the moving range of the vehicle recognition frame obtained by vehicle detection exceeds the moving threshold, it is determined that the vehicle is in a moving state; when the vehicle is in a moving state, and the average vehicle speed, current vehicle quantity, and lane occupancy all exceed the corresponding congestion thresholds, it is determined to be in a congested state;

[0013] An abnormal state output module, configured to send real-time information and log information when it is determined that an abnormal event occurs, and the abnormal events include congestion and illegal parking.

[0014] Further, the data acquisition module pulls the video streams of the monitoring cameras through the RTSP protocol.

[0015] Further, the vehicle data storage module calculates each index by using a time sliding window and a queue caching mechanism.

[0016] Further, the real-time information sent by the abnormal state output module includes the congestion occurrence time, vehicle quantity, and average speed;

[0017] The log information includes pictures, the channels to which the pictures belong, and the positions of illegally parked vehicles.

[0018] Further, the log information is sent through the TCP protocol and uses a packet splitting and retransmission mechanism;

[0019] The picture data is compressed using JPEG.

[0020] The present invention also provides a method for status monitoring and alarm for the entrance and exit of the basement, including the following steps:

[0021] Collect the underground garage scene, train the vehicle detection module, and calibrate the key monitoring areas of the basement entrance and exit;

[0022] Pull the video streams of each monitoring camera set at the entrance and exit of the underground garage in parallel;

[0023] Perform vehicle detection through the trained vehicle detection module, match the detection results of the current frame with the previous frame trajectory through the Hungarian matching algorithm, and assign a unique tracking ID;

[0024] Calculate and store the unique tracking ID, location, and stay time of the vehicle through the hash table structure;

[0025] According to the data of the stored hash table structure, statistically calculate the vehicle density, vehicle speed, relative delay, current vehicle quantity, and lane occupancy; when the moving range of the vehicle recognition frame obtained by vehicle detection is within the corresponding movement threshold, it is determined that the vehicle is in a stationary state; if the duration of the vehicle in the stationary state is greater than the corresponding stay threshold, it is determined that the current vehicle is illegally parked; when the moving range of the vehicle recognition frame obtained by vehicle detection exceeds the movement threshold, it is determined that the vehicle is in a moving state; when the vehicle is in a moving state and the average vehicle speed, current vehicle quantity, and lane occupancy all exceed the corresponding congestion thresholds, it is determined as a congested state;

[0026] When it is determined that an abnormal event occurs, send real-time information and log information, and the abnormal events include congestion and illegal parking.

[0027] Further, the method pulls the video stream of the monitoring camera through the RTSP protocol.

[0028] Further, the method calculates each index by using a time sliding window and a queue caching mechanism.

[0029] Further, the real-time information sent by the method includes the congestion occurrence time, vehicle quantity, and average speed;

[0030] The log information includes pictures, the channels to which the pictures belong, and the locations of illegally parked vehicles.

[0031] Further, the log information is sent using the TCP protocol and adopts a packet splitting and retransmission mechanism;

[0032] The picture data is compressed using JPEG.

[0033] Compared with the prior art, the present invention has the following advantages:

[0034] (1) In terms of cost, the present invention uses existing surveillance cameras, which only need to have network streaming capabilities, without the need to install additional dedicated sensors or perform large-scale modifications, significantly reducing the deployment cost of the system. In addition, through the efficient use of multi-threading technology and GPU acceleration technology, a single edge computing device can support real-time processing of up to 15 video streams, further reducing the deployment and maintenance costs of hardware.

[0035] In terms of system performance, the present invention integrates advanced technologies such as computer vision and edge computing. The system can quickly identify and accurately judge abnormal events such as congestion and illegal parking, effectively improving the monitoring response speed and detection accuracy. It also transmits data in real time through the UDP protocol to ensure rapid early warning of abnormal events.

[0036] (2) At the same time, the present invention realizes stable storage and transmission of log information through the TCP protocol for subsequent data analysis. In addition, the system comprehensively reflects the traffic status of the entrance through multiple indicators, avoiding the deviation caused by the judgment of a single indicator, and further improving the accuracy and reliability of abnormal event detection. In terms of traffic management efficiency, the system outputs accurate traffic status information in real time, providing managers with a scientific basis for decision-making, and can quickly implement guidance and diversion measures, effectively alleviate the congestion of the entrances and exits of the yard, and improve the management efficiency of the parking lot. At the same time, it reduces the interference of abnormal events on the traffic order of surrounding roads and optimizes the traffic smoothness of the overall area. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic flow chart of a method for monitoring and alarming a state of a basement entrance and exit provided in an embodiment of the present invention;

[0038] Figure 2 A schematic diagram of a data preprocessing and marking ROI area provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0040] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0042] Example 1

[0043] This embodiment provides a status monitoring and alarm system for the basement entrance and exit. The system uses a camera as a core detection element to perform all-round monitoring of the main road around the exit and the gate entrance area. The specific steps include:

[0044] The pre-processing module is used to collect underground parking scenes, train the vehicle detection module, and calibrate the key monitoring areas of the underground parking entrances and exits;

[0045] The data acquisition module is used to access multiple surveillance cameras and pull data from the video streams of each surveillance camera in parallel;

[0046] The edge computing unit is used to detect vehicles using the trained vehicle detection module, match the detection result of the current frame with the trajectory of the previous frame using the Hungarian matching algorithm, and assign a unique tracking ID;

[0047] A vehicle data storage module, used to calculate and store the unique tracking ID, location and stay time of the vehicle through a hash table structure;

[0048] The abnormality judgment module is used to count the vehicle density, vehicle speed, relative delay, current number of vehicles and number of lanes occupied according to the stored hash table structure data; when the moving range of the vehicle identification frame obtained by vehicle detection is within the corresponding moving threshold, the vehicle is judged to be in a stationary state; if the duration of the vehicle being in a stationary state is greater than the corresponding stay threshold, the current vehicle is judged to be illegally parked; when the moving range of the vehicle identification frame obtained by vehicle detection exceeds the moving threshold, the vehicle is judged to be in a moving state; when the vehicle is in a moving state, and the average vehicle speed, the current number of vehicles and the number of lanes occupied all exceed the corresponding congestion threshold, it is judged to be in a congested state;

[0049] The abnormal status output module is used to send real-time information and log information when abnormal events occur, including congestion and illegal parking.

[0050] Specifically, the processing process of the preprocessing module is as follows: First, collect the underground storage scenario and train the vehicle detection model. Second, use the image coordinate system to calibrate the key monitoring areas of the basement entrance and exit, narrow the calculation range, and improve the processing efficiency. The ROI area is selected by a polygon and calibrated by combining manual annotation and automatic segmentation algorithms.

[0051] The processing process of the data acquisition module is as follows: Based on the RTSP protocol, access multiple monitoring cameras. A single edge computing device can support parallel processing of up to 15 RTSP video streams at most. Through the caching mechanism and multi-threaded distribution technology, the latency and stability of the video stream are optimized.

[0052] The processing process of the edge computing unit is as follows: First, the recognition and tracking module: Use the trained model as a vehicle detector, and use GPU acceleration for real-time inference. Use the Kalman filter to predict the target position. Match the detection results of the current frame with the trajectory of the previous frame through the Hungarian matching algorithm, and assign a unique tracking ID.

[0053] The processing process of the vehicle data storage module is as follows: Design a hash table structure to store data such as the ID, location, and residence time of vehicles in a short period of time, which is convenient for quick query and update. During the calculation process of each index, a time sliding window and a queue caching mechanism are adopted to reduce redundant calculations.

[0054] The processing process of the abnormal judgment module is as follows: Through the stored hash table data, count the vehicle density, vehicle speed, relative delay, current vehicle quantity, lane occupancy, etc. in the monitoring area. When the moving range of the vehicle recognition frame is within the set threshold, it is judged that the vehicle is in a stationary state. If the stationary state duration is greater than the set threshold, it is judged as illegal parking. If it is in a moving state and indicators such as the average vehicle speed, current vehicle quantity, and lane occupancy meet the set threshold, it is judged as congested.

[0055] The processing process of the abnormal state output module is as follows: Abnormal event information is divided into congestion and illegal parking. When an abnormal event occurs, real-time information and log information are sent. The real-time information is sent through the UDP protocol to ensure fast broadcast to the monitoring center and the user terminal. The information includes key parameters such as the congestion occurrence time, vehicle quantity, average speed, etc. The log information contains pictures, the affiliated channel, the location of the illegally parked vehicle, etc. The data volume is large and the real-time requirement is not high, so the TCP protocol is used for sending. During the transmission process, the image data is compressed by JPEG to balance the network bandwidth and transmission delay. And a packet splitting and retransmission mechanism is adopted to ensure the stability and accuracy of TCP transmission.

[0056] As Figure 1 shown, this embodiment also provides a status monitoring and alarm method for the basement entrance and exit, including the following steps:

[0057] S1: Collect the underground garage scenario, train the vehicle detection module, and calibrate the key monitoring areas at the entrance and exit of the garage.

[0058] S2: Pull the video streams of each monitoring camera set at the entrance and exit of the underground garage in parallel.

[0059] S3: Perform vehicle detection through the trained vehicle detection module, match the detection results of the current frame with the previous frame's trajectory through the Hungarian matching algorithm, and assign a unique tracking ID.

[0060] S4: Calculate and store the unique tracking ID, location, and stay time of the vehicle through a hash table structure.

[0061] S5: Statistically calculate the vehicle density, vehicle speed, relative delay, current number of vehicles, and lane occupancy based on the data stored in the hash table structure; when the moving range of the vehicle recognition frame obtained by vehicle detection is within the corresponding movement threshold, it is determined that the vehicle is in a stationary state; if the duration of the vehicle in a stationary state is greater than the corresponding stay threshold, it is determined that the current vehicle is illegally parked; when the moving range of the vehicle recognition frame obtained by vehicle detection exceeds the movement threshold, it is determined that the vehicle is in a moving state; when the vehicle is in a moving state and both the average vehicle speed, the current number of vehicles, and the lane occupancy exceed the corresponding congestion thresholds, it is determined to be in a congested state.

[0062] S6: When an abnormal event is determined to occur, send real-time information and log information, where the abnormal events include congestion and illegal parking.

[0063] In this embodiment, the processing process of the above method is specifically as follows:

[0064] Obtain the required parameters through the preprocessing step, such as the pixel vertices of the ROI polygon area and the monitored road length L, as Figure 2 shown. During the operation of the system of the present invention, first pull the video stream through the RTSP protocol, and use the OpenCV library for decoding and frame rate synchronization. Optimize the delay and stability of the video stream through the cache mechanism and multi-threaded distribution technology. Use the trained model as the vehicle detector and utilize GPU acceleration for real-time inference. Use the Kalman filter to predict the target location. Match the detection results of the current frame with the previous frame's trajectory through the Hungarian matching algorithm and assign a unique tracking ID.

[0065] In the vehicle data storage part, store data such as the ID, location, and stay time of the vehicles present within 2 minutes through a hash table structure, which is convenient for quick query and update. During the calculation of each index, adopt a time sliding window and queue cache mechanism to statistically calculate the vehicle density, vehicle speed, relative delay, current number of vehicles, lane occupancy, etc. in the monitoring area, and reduce redundant calculations.

[0066] When the moving range of the vehicle recognition frame |x 1 -x 2 |<0.2w and |y 1 -y 2 |<0.2h are both true, it is determined that the vehicle is in a stationary state. If the stationary state duration is greater than the set threshold of 2 minutes, it is determined as illegal parking. When the following indicators: vehicle average speed 0m / s < v < 1m / s, vehicle maximum speed v max <1, current vehicle count car_current > 1, lane occupancy lane_occupancy > 1, relative delay Relative_delay >= 0 are all satisfied, it is determined as congested. Once congestion or illegal parking occurs, key parameters such as the current time, event type, vehicle quantity, average speed, etc. will be sent to the server through the UDP protocol, and the server will update the user-side applet and the management-side induction screen in real time. At the same time, the current screen will be compressed using JPEG to balance network bandwidth and transmission delay. And a packet splitting and retransmission mechanism will be adopted to ensure the stability and accuracy of TCP transmission of image data.

[0067] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations based on the concept of the present invention without creative labor. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field of the present invention through logical analysis, reasoning, or limited experiments based on the concept of the present invention on the basis of the prior art should be within the protection scope determined by the claims.

Claims

1. A status monitoring and alarm system for basement entrances and exits, characterized in that: include: The pre-processing module is used to collect underground parking scenes, train the vehicle detection module, and calibrate the key monitoring areas of the underground parking entrances and exits; The data acquisition module is used to access multiple surveillance cameras and pull data from the video streams of each surveillance camera in parallel; The edge computing unit is used to detect vehicles using the trained vehicle detection module, match the detection result of the current frame with the trajectory of the previous frame using the Hungarian matching algorithm, and assign a unique tracking ID; A vehicle data storage module, used to calculate and store the unique tracking ID, location and stay time of the vehicle through a hash table structure; The abnormality judgment module is used to count the vehicle density, vehicle speed, relative delay, current number of vehicles and number of lane occupancy according to the stored hash table structure data; when the moving range of the vehicle identification frame obtained by vehicle detection is within the corresponding moving threshold, it is judged that the vehicle is in a stationary state; If the time the vehicle is stationary is longer than the corresponding stop threshold, the vehicle is judged to be illegally parked; when the moving range of the vehicle identification frame obtained by vehicle detection exceeds the moving threshold, the vehicle is judged to be in motion; when the vehicle is in motion and the average vehicle speed, the current number of vehicles and the number of lanes occupied all exceed the corresponding congestion threshold, it is judged to be in a congested state; The abnormal status output module is used to send real-time information and log information when it is determined that an abnormal event occurs, and the abnormal event includes congestion and illegal parking.

2. A status monitoring and alarm system for basement entrances and exits according to claim 1, characterized in that: The data acquisition module pulls the video stream of the surveillance camera through the RTSP protocol.

3. A status monitoring and alarm system for basement entrances and exits according to claim 1, characterized in that: The vehicle data storage module uses a time sliding window and queue cache mechanism to calculate various indicators.

4. A status monitoring and alarm system for basement entrances and exits according to claim 1, characterized in that: The real-time information sent by the abnormal state output module includes the time of congestion occurrence, the number of vehicles and the average speed; The log information includes the picture, the channel to which the picture belongs, and the location of the illegally parked vehicle.

5. A status monitoring and alarm system for basement entrances and exits according to claim 4, characterized in that: The log information is sent using the TCP protocol, and uses a packetization and retransmission mechanism; The image data is compressed using JPEG.

6. A method for monitoring and alarming the status of a basement entrance and exit, characterized in that: The following steps are involved: Collect underground parking scenes, train vehicle detection modules, and calibrate key monitoring areas at the entrances and exits of underground parking lots; Parallel data extraction of video streams from various surveillance cameras installed at the entrances and exits of underground warehouses; The trained vehicle detection module is used to detect vehicles, and the detection result of the current frame is matched with the trajectory of the previous frame through the Hungarian matching algorithm, and a unique tracking ID is assigned; The unique tracking ID, location and dwell time of the vehicle are calculated and stored through a hash table structure; The vehicle density, vehicle speed, relative delay, current number of vehicles and lane occupancy are counted according to the stored hash table structure data; when the moving range of the vehicle identification frame obtained by vehicle detection is within the corresponding moving threshold, it is judged that the vehicle is stationary; If the time the vehicle is stationary is longer than the corresponding stop threshold, the vehicle is judged to be illegally parked; when the moving range of the vehicle identification frame obtained by vehicle detection exceeds the moving threshold, the vehicle is judged to be in motion; when the vehicle is in motion and the average vehicle speed, the current number of vehicles and the number of lanes occupied all exceed the corresponding congestion threshold, it is judged to be in a congested state; When it is determined that an abnormal event occurs, real-time information and log information are sent, and the abnormal event includes congestion and illegal parking.

7. A method for monitoring and alarming the status of a basement entrance and exit according to claim 6, characterized in that: The method uses the RTSP protocol to pull the video stream of the surveillance camera.

8. A method for monitoring and alarming the status of a basement entrance and exit according to claim 6, characterized in that: The method adopts a time sliding window and a queue cache mechanism to calculate various indicators.

9. A method for monitoring and alarming the status of a basement entrance and exit according to claim 6, characterized in that: The real-time information sent by the method includes the time when the congestion occurs, the number of vehicles and the average speed; The log information includes the picture, the channel to which the picture belongs, and the location of the illegally parked vehicle.

10. A method for monitoring and alarming the status of a basement entrance and exit according to claim 9, characterized in that: The log information is sent using the TCP protocol, and uses a packetization and retransmission mechanism; The image data is compressed using JPEG.

Images