A video analysis technology-based coal pile left in a carbody early warning system

Abstract

The present application relates to the technical field of video analysis, in particular to a kind of based on video analysis technology's car compartment coal pile left over early warning system, it includes front end, weighbridge system server and client;Analysis host carries out car compartment detection tracking to the video sequence of network camera access, when judging car compartment stationary, carry out tarpaulin covering detection, if tarpaulin covering is detected, then directly produce alarm;If tarpaulin covering is not detected, then use coal pile detection module to carry out coal pile left over detection;If coal pile is detected, then produce alarm;The alarm information generated is pushed to weighbridge system server;Weighbridge system server receives the alarm information pushed by analysis host, and controls the opening and closing of brake, simultaneously, alarm record is stored;Client receives alarm information in real time, and historical alarm data can be inquired.The present application comprehensively uses car compartment detection tracking, tarpaulin detection and coal pile detection technology, solves the problem of coal loss in factory caused by the accuracy performance problem of weighbridge system.

Description

Technical Field

[0001] This invention relates to the field of video analytics technology, and in particular to an early warning system for coal piles left in railway carriages based on video analytics technology. Background Technology

[0002] There are instances of trucks transporting coal to the coking plant carrying small amounts of coal out of the plant without authorization. Specifically, trucks that have just unloaded coal pass through a weighbridge before leaving the plant. During the weighing process, due to inaccurate readings and other reasons, the weighbridge adjusts the weighing threshold to ±200KG, meaning that coal weighing less than 200KG is not considered overweight.

[0003] For the reasons mentioned above, after trucks unload coal at the plant, sometimes some coal, weighing up to 200 kg, remains inside the truck. To prevent this from happening, a visual algorithm is proposed to identify such behavior and issue timely warnings to prevent this problem from recurring. Summary of the Invention

[0004] The purpose of this invention is to address the problems existing in the background technology by proposing a car coal pile loss early warning system based on video analysis technology. By using car detection and tracking, tarpaulin detection and coal pile detection technologies, the system solves the problem of coal loss in the plant caused by the inaccuracy of the weighbridge system.

[0005] The technical solution of the present invention is an early warning system for coal pile residue in a freight car based on video analysis technology, comprising a front end, a weighbridge system server, and a client.

[0006] Data transmission between the front-end, the weighbridge system server, and the client is conducted via a switch;

[0007] The front end includes a network camera and an analysis host;

[0008] Network cameras capture real-time video of the monitored area and send it to the analysis host via a switch;

[0009] The main analysis unit includes a carriage detection and tracking module, a tarpaulin coverage detection module, a carriage stationary filtration module, and a coal pile detection module;

[0010] The analysis host performs carriage detection and tracking on the video sequences accessed by the network camera. When it determines that the carriage is stationary, it performs tarpaulin coverage detection. If tarpaulin coverage is detected, an alarm is generated directly.

[0011] If no tarpaulin coverage is detected, the coal pile detection module is used to detect any remaining coal piles; if a coal pile is detected, an alarm is generated; the generated alarm information is pushed to the weighbridge system server.

[0012] The weighbridge system server receives and analyzes the alarm information pushed by the host, controls the opening and closing of the vehicle gate, and stores the alarm records.

[0013] The client receives alarm information in real time and can query historical alarm data.

[0014] The carriage detection and tracking module analyzes the host computer to monitor and track the carriages in the camera video stream in real time, generating real-time coordinates and trajectory data of the carriages.

[0015] Preferably, the carriage detection and tracking module processes 7 frames per second.

[0016] The vehicle stationary filtering module judges the overlap range of two adjacent frames. If the overlap range is greater than or equal to the set threshold and exceeds the set number of times, it is judged that the vehicle is stationary and the subsequent detection of the tarpaulin on the vehicle is performed.

[0017] If the overlap between two adjacent frames is less than the threshold, then return to carriage detection and tracking.

[0018] Preferably, the overlap range is 0.1-1, with a default setting of 0.5; the number of times is set to 1-10, with a default setting of 5.

[0019] The tarpaulin coverage detection module includes a tarpaulin layout reliability detection unit and a tarpaulin size judgment unit;

[0020] The tent arrangement reliability detection unit performs multiple tests on the tent. If the results of multiple tests are all greater than or equal to the set threshold, the size of the tent is judged; if the results are all less than the threshold, the subsequent coal pile test is performed. The number of tests ranges from 1 to 10, with a default of 5.

[0021] The tarpaulin size judgment unit detects the tarpaulin size. If the tarpaulin size is greater than or equal to the threshold, it pushes the full alarm image and alarm video clip to the weighbridge system server.

[0022] If the tarpaulin size is smaller than the threshold, a coal pile test is performed.

[0023] The tarpaulin size ranges from 100 to 1000, with a default value of 100. A value of 100 indicates a pixel size of 100*100. If the tarpaulin size is greater than 100*100 pixels, it is considered to meet the testing requirements.

[0024] The coal pile detection module includes a coal pile confidence detection unit and a coal pile size detection unit;

[0025] The coal pile confidence detection unit performs multiple tests on the coal pile. If the results of multiple tests are greater than or equal to the set threshold, the size of the coal pile is judged. If the results are less than the threshold multiple times, images and video clips are pushed to the weighbridge system server. The judgment value range is 1-10 times, with a default of 5 times.

[0026] The coal pile size detection unit detects the size of the coal pile. If the coal pile size is greater than or equal to the threshold, it pushes the full alarm image and alarm video clip to the weighbridge system server; if the coal pile size is less than the threshold, it pushes the image and video clip to the weighbridge system server.

[0027] The coal pile size ranges from 60 to 1000, with a default value of 60. If the value is 60, it means that the size in pixels is 60*60. If the coal pile size in pixels is greater than 60*60, it can be considered to meet the detection requirements.

[0028] Preferably, the tent layout confidence detection unit and the coal pile confidence detection unit establish an identification model based on the target detection algorithm, and improve the identification rate of the identification model through data collection and data training.

[0029] Compared with the prior art, the present invention has the following beneficial technical effects:

[0030] 1. This invention uses multiple network cameras to collect video in real time. The carriage detection and tracking module analyzes the host to monitor and track the carriages in the camera video stream in real time, generating real-time coordinates and trajectory data of the carriages. The vehicle is first judged to be stationary, which provides a stable video image for subsequent tarpaulin detection and coal pile detection, ensuring the accuracy of the algorithm and reducing false alarms.

[0031] 2. This invention utilizes vehicle compartment detection and tracking, tarpaulin detection, and coal pile detection technologies to perform confidence and size detection on the tarpaulin and coal pile respectively. Based on the detection results, it determines whether a coal pile exists and controls the opening and closing of the vehicle gate through the weighbridge system server, and pushes alarm information; ultimately solving the problem of coal loss in the plant caused by the accuracy and performance issues of the weighbridge system. Attached Figure Description

[0032] Figure 1 This is a system topology diagram of the early warning system of the present invention;

[0033] Figure 2 This is a flowchart of the early warning system of the present invention. Detailed Implementation

[0034] Example 1

[0035] like Figure 1 As shown, the present invention proposes a video analytics-based early warning system for coal piles left in freight cars, comprising a front-end, a weighbridge system server, and a client.

[0036] Data transmission between the front-end, the weighbridge system server, and the client is conducted via a switch;

[0037] The front end includes a network camera and an analysis host. The camera is a wide-angle camera, and its field of view can cover the entire carriage.

[0038] Network cameras capture real-time video of the monitored area and send it to the analysis host via a switch;

[0039] The main analysis unit includes a carriage detection and tracking module, a tarpaulin coverage detection module, a carriage stationary filtration module, and a coal pile detection module;

[0040] The analysis host performs carriage detection and tracking on the video sequences accessed by the network camera. When it determines that the carriage is stationary, it performs tarpaulin coverage detection. If tarpaulin coverage is detected, an alarm is generated directly.

[0041] If no tarpaulin coverage is detected, the coal pile detection module is used to detect any remaining coal piles; if a coal pile is detected, an alarm is generated; the generated alarm information is pushed to the weighbridge system server.

[0042] The weighbridge system server receives and analyzes the alarm information pushed by the host, controls the opening and closing of the vehicle gate, and stores the alarm records.

[0043] The client receives alarm information in real time and can query historical alarm data.

[0044] This invention integrates carriage detection and tracking, tarpaulin detection, and coal pile detection technologies to solve the problem of coal loss within the plant caused by the accuracy and performance issues of the weighbridge system.

[0045] like Figure 2 As shown, the workflow of the early warning system for coal pile residue in a freight car based on video analytics proposed in this invention is as follows:

[0046] 1. A camera captures real-time video of the monitored area;

[0047] 2. The carriage detection and tracking module monitors and tracks the carriages in the camera video stream in real time through the analysis host, generating real-time carriage coordinates and trajectory data; to reduce the computational load on the analysis host, the carriage detection and tracking module processes 7 frames per second.

[0048] 3. The carriage stationary filtering module judges the overlap range of two adjacent frames. If the overlap range is greater than or equal to the set threshold and exceeds the set number of times, it is judged that the vehicle is stationary and the subsequent detection of the tarpaulin on the carriage is performed.

[0049] If the overlap between two adjacent frames is less than the threshold, then return to carriage detection and tracking. The overlap range is 0.1-1, with a default of 0.5; the number of times is set is 1-10, with a default of 5.

[0050] 4. The tarpaulin coverage detection module's tarpaulin layout reliability detection unit performs multiple tests on the tarpaulin. If the results of multiple tests are all greater than or equal to the set threshold, the tarpaulin size is judged; if the results are all less than the threshold, subsequent coal pile detection is performed. The number of tests ranges from 1 to 10, with a default of 5.

[0051] 5. The tarpaulin size judgment unit in the tarpaulin coverage detection module detects the tarpaulin size. If the tarpaulin size is greater than or equal to the threshold, the alarm full image and alarm video clip are pushed to the weighbridge system server.

[0052] If the tarpaulin size is smaller than the threshold, a coal pile test is performed.

[0053] The tarpaulin size ranges from 100 to 1000, with a default of 100. If the value is 100, it means that the size in pixels is 100*100. If the tarpaulin size in pixels is greater than 100*100, it can be considered to meet the testing requirements.

[0054] 6. The coal pile confidence detection unit in the coal pile detection module performs multiple tests on the coal pile. If the results of multiple tests are greater than or equal to the set threshold, the size of the coal pile is judged. If the results are less than the threshold multiple times, images and video clips are pushed to the weighbridge system server. The judgment value range is 1-10 times, with a default of 5 times.

[0055] 7. The coal pile size detection unit in the coal pile detection module detects the size of the coal pile. If the coal pile size is greater than or equal to the threshold, an alarm full image and alarm video clip are pushed to the weighbridge system server; if the coal pile size is less than the threshold, an image and video clip are pushed to the weighbridge system server.

[0056] The coal pile size ranges from 60 to 1000, with a default value of 60. A value of 60 indicates a size of 60*60 pixels. If the coal pile size is greater than 60*60 pixels, it is considered to meet the testing requirements.

[0057] 8. The weighbridge system server receives alarm information from the intelligent analysis host and performs gate opening or closing operations. It also stores full alarm images and alarm video clips for event retrospective retrieval.

[0058] 9. The client can display the full alarm graph and alarm video in real time, and can query historical records.

[0059] In this embodiment, a network camera captures video in real time. The carriage detection and tracking module analyzes the host to monitor and track the carriages in the camera's video stream in real time, generating real-time coordinates and trajectory data of the carriages. First, the vehicle is judged to be stationary, providing a stable video image for subsequent tarpaulin detection and coal pile detection, which can ensure the accuracy of the algorithm and reduce false alarms. Then, by using carriage detection and tracking, tarpaulin detection and coal pile detection technologies, the problem of coal loss in the plant caused by the accuracy and performance issues of the weighbridge system is solved.

[0060] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A video analytics-based early warning system for coal pile residue in railway carriages, characterized in that, This includes the front-end, weighbridge system server, and client; Data transmission between the front-end, the weighbridge system server, and the client is conducted via a switch; The front end includes a network camera and an analysis host; Network cameras capture real-time video of the monitored area and send it to the analysis host via a switch; The main analysis unit includes a carriage detection and tracking module, a tarpaulin coverage detection module, a carriage stationary filtration module, and a coal pile detection module; The analysis host performs carriage detection and tracking on the video sequences accessed by the network camera. When it determines that the carriage is stationary, it performs tarpaulin coverage detection. If tarpaulin coverage is detected, an alarm is generated directly. If no tarpaulin coverage is detected, the coal pile detection module is used to detect any remaining coal piles; if a coal pile is detected, an alarm is generated; the generated alarm information is pushed to the weighbridge system server. The weighbridge system server receives and analyzes alarm information pushed by the host, controls the opening and closing of the vehicle gate, and stores alarm records. The client receives alarm information in real time and can query historical alarm data.

2. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 1, characterized in that, The carriage detection and tracking module analyzes the host computer to monitor and track the carriages in the camera video stream in real time, generating real-time coordinates and trajectory data of the carriages.

3. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 2, characterized in that, The carriage detection and tracking module processes data at a rate of 7 frames per second.

4. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 1, characterized in that, The vehicle stationary filtering module judges the overlap range of two adjacent frames. If the overlap range is greater than or equal to the set threshold and exceeds the set number of times, it is judged that the vehicle is stationary and the subsequent detection of the tarpaulin on the vehicle is performed. If the overlap between two adjacent frames is less than the threshold, then return to carriage detection and tracking.

5. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 4, characterized in that, The overlap range is 0.1-1; the number of overlaps is 1-10.

6. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 1, characterized in that, The tarpaulin coverage detection module includes a tarpaulin layout reliability detection unit and a tarpaulin size judgment unit; The tent arrangement reliability detection unit performs multiple tests on the tent. If the results of multiple tests are all greater than or equal to the set threshold, the size of the tent is judged; if the results are all less than the threshold, the coal pile is then tested. The number of tests ranges from 1 to 10. The coal pile detection module includes a coal pile confidence detection unit and a coal pile size detection unit; The coal pile confidence detection unit performs multiple tests on the coal pile. If the results of multiple tests are greater than or equal to the set threshold, the size of the coal pile is determined. If the results are less than the threshold, images and video clips are pushed to the weighbridge system server. The range of values ​​to be judged is 1-10.

7. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 6, characterized in that, The tarpaulin size judgment unit detects the tarpaulin size. If the tarpaulin size is greater than or equal to the threshold, it pushes the full alarm image and alarm video clip to the weighbridge system server. If the tarpaulin size is smaller than the threshold, a coal pile test is performed. The size of the tarpaulin ranges from 100 to 1000.

8. The early warning system for coal pile residue in a freight car based on video analytics technology according to claim 6, characterized in that, The coal pile size detection unit detects the size of the coal pile. If the coal pile size is greater than or equal to the threshold, it pushes the full alarm image and alarm video clip to the weighbridge system server; if the coal pile size is less than the threshold, it pushes the image and video clip to the weighbridge system server. The coal pile size ranges from 60 to 1000.

9. A video analytics-based early warning system for coal pile residue in a freight car, as described in claim 6, is characterized in that... The tent layout confidence detection unit and the coal pile confidence detection unit establish recognition models based on target detection algorithms, and improve the recognition rate of the recognition models through data collection and data training.

Images