Sea area management and control early warning evidence obtaining method based on video intelligent analysis

Abstract

The invention relates to a sea area management and control early warning evidence obtaining method based on video intelligent analysis, and the method comprises the steps: collecting equipment parameters: collecting internal and external parameters of equipment and environment parameters; equipment screening control is carried out, controllable and controllable equipment with a good detection effect is selected through equipment state analysis and target detection effectiveness analysis, and follow-up monitoring of a target area is realized through pan-tilt control; performing video intelligent analysis, determining a follow-up concerned target through mapping matching of a video intelligent analysis model and different coordinate system positioning information, and performing accurate linkage and video snapshot by fine tuning equipment; according to early warning and evidence obtaining association, structured attribute analysis is performed on a target and an area detected in a video capturing process, and an evidence obtaining result is added to early warning information with an association relationship, so that repeated evidence obtaining is avoided. According to the invention, the accuracy and efficiency of early warning evidence obtaining can be effectively improved, the utilization efficiency of data and resources is improved, and powerful technical support is provided for sea area management and control.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital image processing and relates to a sea area control early warning and evidence collection method based on video intelligent analysis. When the early warning is triggered, the target is locked and video is captured through device linkage control, deep learning, and association matching algorithms to achieve intelligent evidence collection and analysis of the early warning target. Background Art

[0002] With the rapid development of the marine economy, the importance of sea area control has become increasingly prominent. Sea area control requires timely discovery and handling of illegal fishing, maritime smuggling, illegal navigation and other illegal activities. Therefore, in sea area control, timely and accurate analysis and handling of sea area warning targets is the key.

[0003] Although the existing sea area control system has achieved a certain degree of information management, it still faces many challenges, such as weak on-site perception, old camera equipment, and poor communication signals, which make it difficult to obtain positioning data for the target. In response to these problems, the industry has proposed a variety of early warning and evidence collection technology solutions, which integrate video surveillance, image recognition and other technologies to achieve automatic detection and recording of early warning targets. Among them, automatic evidence collection technology is particularly eye-catching. The achievements in automatic evidence collection include the intelligent supervision system of electronic boundary stakes, which is linked with traditional security products when unknown threats are discovered, and can realize intelligent identification and early warning push of suspected illegal sea use behaviors.

[0004] However, these solutions still have many shortcomings in practical applications, such as low recognition accuracy, slow response speed, and inability to adapt to complex and changing marine environments. In addition, when the existing early warning system is triggered, multiple other types of early warning intelligence will be generated at the same time. Each early warning intelligence will independently call the device for targeted capture and evidence collection, which increases the cost and energy consumption of the system and may interfere with the normal business of on-duty personnel. Summary of the invention

[0005] The purpose of the present invention is to address the above shortcomings and propose a sea area control early warning and evidence collection method based on video intelligent analysis to solve the problems of the existing automatic early warning and evidence collection being unable to take into account environmental change factors, inaccurate evidence collection results, and repeated evidence collection caused by the simultaneous triggering of multiple warnings, so as to assist on-duty personnel to better confirm the authenticity of the current early warning intelligence and carry out accurate and efficient case handling.

[0006] The technical solution of the present invention is as follows:

[0007] A method for collecting evidence for sea area control early warning based on video intelligent analysis, characterized in that it comprises the following steps:

[0008] (1) Equipment parameter collection: Collect the internal and external parameters of the equipment and the environmental parameters to lay the foundation for accurate control of the equipment;

[0009] (2) Equipment screening and control: Through equipment status analysis and target detection effectiveness analysis, controllable and effective equipment is selected, and then the target area is monitored dynamically through pan-tilt control;

[0010] (3) Video intelligent analysis: Through the video intelligent analysis model and the mapping and matching of positioning information in different coordinate systems, the target of interest is determined and the equipment is fine-tuned for precise linkage and video capture;

[0011] (4) Early warning and evidence association: By performing structured attribute analysis on the targets and areas detected during the video capture process, the evidence collection results are added to the early warning intelligence with associated relationships, avoiding duplicate evidence collection and improving the efficiency of evidence utilization.

[0012] In the above technical solution, the equipment parameter collection includes equipment calibration and visual field calibration;

[0013] The device calibration first collects the external parameters of the device, including the positioning position, altitude, installation level quality, and true north deflection angle, and obtains the internal parameters of the device through the model notes of the device, including the field of view angle change relationship and angle range, and then verifies and calibrates the internal and external parameters of the device through calibration target data collection, and finally obtains the accurate internal and external parameters of the device;

[0014] The field of view calibration is achieved by controlling the device's fan-scan inspection to obtain the device's observation range and the obstructions in each direction, and determine the device's effective detection area at each horizontal angle. During the calibration process, a circle of 8 images is recorded at intervals of 45 degrees, and compared with the image algorithm in subsequent regular inspections to determine whether the device has deviated and issue an alarm in time to ensure the accuracy of the recorded parameters.

[0015] In the above technical solution, the device screening control includes evidence collection decision, screening of valid devices, secondary screening of valid visual areas and device linkage control;

[0016] The evidence collection decision is made by comparing the currently executed evidence collection threads. When the same target and the same area exist, no new evidence collection process will be performed, thereby avoiding repeated evidence collection and improving evidence collection efficiency.

[0017] The effective device screening first excludes devices beyond the visual range by calculating the straight-line distance between each device and the target, then selects devices with better linkage effects according to the control quality sorting, and finally screens idle devices that meet the requirements according to the current occupancy status of the devices;

[0018] The effective visual field secondary screening is based on the device visual field calibration to determine whether the current target is in the blind spot of the candidate device, and finally determine the optimal device that can detect the target;

[0019] The device linkage control calculates the polar coordinates of the target relative to the device, determines the device control parameters through coordinate system conversion operations, and the linkage device is continuously controlled with a field of view angle of a set proportion.

[0020] In the above technical solution, the video intelligent analysis includes target detection and tracking, positioning information mapping and matching, target locking and video capture;

[0021] The target detection and tracking performs intelligent analysis on the video based on a deep learning model to determine the positions of multiple targets in the video and obtain the motion relationship of each target through a multi-target tracking and matching algorithm;

[0022] The positioning information mapping and matching is to determine the identity information of the target by performing cross-modal conversion on the radar and AIS target positioning information within the field of view, mapping it to the pixel coordinate system of the video screen, and associating and matching it with the detected target;

[0023] The target locking is to determine the specific position of the linkage target in the video according to its identity information, and control the device to follow the target in real time according to the actual situation of the target in the video, so as to achieve the target locking and precise linkage;

[0024] The video capture is to collect evidence records of the video images after the current target is locked, and obtain a 10-second video or snapshot to assist the later duty personnel in analysis and confirmation.

[0025] In the above technical solution, the early warning evidence association includes video target structuring, multi-label storage, and related intelligence association multiplexing;

[0026] The video target structuring is to detect and extract attributes of all targets in the video or snapshot of the evidence collection, including the identity of the target, the target area currently detected, and the suspected abnormal behavior, so as to fully mine the important information in the evidence collection results;

[0027] The multi-label storage, based on the video target structured analysis results, establishes corresponding labels for different attributes, stores them by category in the database, and realizes query screening in the comprehensive query system;

[0028] The related intelligence association multiplexing is based on the results of the video target structured analysis, and is associated and matched with other warning intelligence triggered in the same period. The access address of the forensic result is added to the warning intelligence with a higher correlation relationship to achieve the reuse of the forensic result.

[0029] Compared with the prior art, the advantages of the present invention are: optimizing the design to address the problems of field of view obstruction, state occupancy, repeated evidence collection, etc. that affect automatic early warning and evidence collection, fully combining intelligent analysis algorithms to realize automatic early warning and evidence collection of equipment, and improving the reuse rate of evidence collection results, so as to achieve efficient and accurate detection, tracking, matching and locking of targets in video images, thereby effectively improving the accuracy and efficiency of early warning and evidence collection, improving the efficiency of data and resource utilization, achieving the expected results in business scenarios, and providing strong technical support for sea area management and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a framework diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0031] like Figure 1 As shown, the present invention includes equipment parameter collection, equipment screening control, video intelligent analysis, and early warning and evidence collection association.

[0032] The equipment parameter collection realizes the collection of internal and external parameters of the equipment and environmental parameters, laying a foundation for accurate control of the equipment;

[0033] The device screening control selects controllable, controllable and good detection effect devices through device status analysis and target detection effectiveness analysis, and then realizes follow-up monitoring of the target area through pan-tilt control;

[0034] The video intelligent analysis determines the follow-up target of interest and fine-tunes the equipment for precise linkage and video capture through the video intelligent analysis model and the mapping and matching of the positioning information of different coordinate systems;

[0035] The early warning evidence association performs structured attribute analysis on the targets and areas detected during the video capture process, and adds the evidence collection results to the early warning intelligence with associated relationships, thereby avoiding repeated evidence collection and improving the efficiency of evidence collection.

[0036] The equipment parameter collection is the basis for the implementation of this technology, including equipment calibration and field of view calibration. The equipment calibration first collects the equipment's external parameters such as the positioning position, altitude, installation level quality, and true north deflection angle through special equipment, and obtains the equipment's field of view angle change relationship, angle range and other internal parameters through the equipment's model notes, and then verifies and calibrates the equipment's internal and external parameters through calibration target data collection, and finally obtains the equipment's accurate internal and external parameters; the field of view calibration obtains the equipment's observation range and the obstructions in each direction by controlling the equipment's fan-scanning inspection, and determines the effective detection area of ​​the equipment at each horizontal angle. Image records can be saved during the calibration process, and compared with the help of image algorithms in subsequent regular inspections to determine whether the equipment has offset and issue an alarm in time to ensure the accuracy of the recorded parameters.

[0037] The device screening control is a necessary path for the implementation of this technology, including the evidence decision module, screening of effective devices, effective field of view secondary screening, and device linkage control. The evidence decision module compares the currently executed evidence threads, and when there are the same target, the same area, etc., no new evidence process will be performed, avoiding repeated evidence collection and improving evidence collection efficiency; the screening of effective devices implements the preliminary screening of devices, first by calculating the straight-line distance between each device and the target to exclude devices that exceed the visual range, and then selects devices with good linkage effects according to the control quality sorting, and finally screens the idle devices that meet the requirements according to the current occupancy status of the device; the effective field of view secondary screening is based on the device field of view calibration, and determines whether the current target is in the blind spot of the candidate device, and finally determines the optimal device that can detect the target; the device linkage control calculates the polar coordinates of the target relative to the device, determines the device control parameters through coordinate system conversion operations, and the linkage device is continuously controlled with a field of view angle of a set proportion.

[0038] The video intelligent analysis is the core part of this technology, including target detection and tracking, positioning information mapping and matching, target locking, and video capture. The target detection and tracking performs intelligent analysis of the video based on the deep learning model, determines the positions of multiple targets in the video screen, and obtains the motion relationship of each target through the multi-target tracking and matching algorithm; the positioning information mapping and matching converts the target positioning information of radar, AIS and other targets within the field of view across modes, maps it to the pixel coordinate system of the video screen, and associates and matches it with the detected target to determine the identity information of the target; the target locking determines the specific position of the linkage target in the video screen according to its identity information, and controls the real-time tracking of the device according to the actual situation of the target in the video, so as to achieve the focus target locking and precise linkage; the video capture realizes the evidence recording of the video screen after the current target is locked through the background service, obtains a 10-second video or snapshot, and assists the later duty personnel in the judgment and confirmation.

[0039] The warning evidence association is a business visualization presentation of this technology, including video target structuring, multi-label storage, and related intelligence association multiplexing. The video target structuring detects and extracts attributes of all targets in the video or snapshot of the evidence, including the identity of the target, the target area currently detected, suspected abnormal behavior, etc., to fully mine the important information in the evidence results; the multi-label storage is based on the results of the video target structured analysis, establishes corresponding labels for different attributes, and stores them in a classified manner in the database, and can realize query screening in the comprehensive query system; the related intelligence association multiplexing is based on the results of the video target structured analysis, and is associated and matched with other warning intelligence triggered in the same period, and the access address of the forensic result is added to the warning intelligence with a higher correlation relationship to realize the reuse of the forensic results.

[0040] In the above technical solution, the process of equipment screening control according to effective range screening, control quality screening, idle state screening, and effective field of view secondary screening is the best implementation path under the current construction environment. Adjusting the screening order during specific implementation can also achieve the same equipment selection effect.

[0041] In the above technical solution, the forensic results of 10-second video recording or snapshots in video intelligent analysis are regular forensic records. During specific implementation, the target's motion trajectory, multi-dimensional feature time series changes, and other lengths of video recording and multi-frame snapshots can also be recorded simultaneously.

Claims

1. A sea area control early warning and evidence collection method based on video intelligent analysis, characterized in that: The following steps are involved: (1) Equipment parameter collection: Collect the internal and external parameters of the equipment and the environmental parameters to lay the foundation for accurate control of the equipment; (2) Equipment screening and control: Through equipment status analysis and target detection effectiveness analysis, controllable and effective equipment is selected, and then the target area is monitored dynamically through pan-tilt control; (3) Video intelligent analysis: Through the video intelligent analysis model and the mapping and matching of positioning information in different coordinate systems, the target of interest is determined and the equipment is fine-tuned for precise linkage and video capture; (4) Early warning and evidence association: By performing structured attribute analysis on the targets and areas detected during the video capture process, the evidence collection results are added to the early warning intelligence with associated relationships, avoiding duplicate evidence collection and improving the efficiency of evidence utilization.

2. The sea area control early warning and evidence collection method based on video intelligent analysis according to claim 1 is characterized by: The equipment parameter collection includes equipment calibration and visual field calibration; The device calibration first collects the external parameters of the device, including the positioning position, altitude, installation level quality, and true north deflection angle, and obtains the internal parameters of the device through the model notes of the device, including the field of view angle change relationship and angle range, and then verifies and calibrates the internal and external parameters of the device through calibration target data collection, and finally obtains the accurate internal and external parameters of the device; The field of view calibration is achieved by controlling the device's fan-scan inspection to obtain the device's observation range and the obstructions in each direction, and determine the device's effective detection area at each horizontal angle. During the calibration process, a circle of 8 images is recorded at intervals of 45 degrees, and compared with the image algorithm in subsequent regular inspections to determine whether the device has deviated and issue an alarm in time to ensure the accuracy of the recorded parameters.

3. The sea area control early warning and evidence collection method based on video intelligent analysis according to claim 1 is characterized by: The device screening control includes evidence collection decision, screening of effective devices, secondary screening of effective visual areas and device linkage control; The evidence collection decision is made by comparing the currently executed evidence collection threads. When the same target and the same area exist, no new evidence collection process will be performed, thereby avoiding repeated evidence collection and improving evidence collection efficiency. The effective device screening first excludes devices beyond the visual range by calculating the straight-line distance between each device and the target, then selects devices with better linkage effects according to the control quality sorting, and finally screens idle devices that meet the requirements according to the current occupancy status of the devices; The effective visual field secondary screening is based on the device visual field calibration to determine whether the current target is in the blind spot of the candidate device, and finally determine the optimal device that can detect the target; The device linkage control calculates the polar coordinates of the target relative to the device, determines the device control parameters through coordinate system conversion operations, and the linkage device is continuously followed and controlled at a set proportional field of view angle.

4. The method for collecting evidence for sea area control and early warning based on video intelligent analysis according to claim 1 is characterized in that: The video intelligent analysis includes target detection and tracking, positioning information mapping and matching, target locking and video capture; The target detection and tracking performs intelligent analysis on the video based on a deep learning model to determine the positions of multiple targets in the video and obtain the motion relationship of each target through a multi-target tracking and matching algorithm; The positioning information mapping and matching is to determine the identity information of the target by performing cross-modal conversion on the radar and AIS target positioning information within the field of view, mapping it to the pixel coordinate system of the video screen, and associating and matching it with the detected target; The target locking is to determine the specific position of the linkage target in the video according to its identity information, and control the device to follow the target in real time according to the actual situation of the target in the video, so as to achieve the target locking and precise linkage; The video capture is to collect evidence records of the video images after the current target is locked, and obtain a 10-second video or snapshot to assist the later duty personnel in analysis and confirmation.

5. The method for collecting evidence for sea area control and early warning based on video intelligent analysis according to claim 1 is characterized in that: The early warning evidence association includes video target structuring, multi-label storage, and related intelligence association multiplexing; The video target structuring is to detect and extract attributes of all targets in the video or snapshot of the evidence collection, including the identity of the target, the target area currently detected, and the suspected abnormal behavior, so as to fully mine the important information in the evidence collection results; The multi-label storage, based on the video target structured analysis results, establishes corresponding labels for different attributes, stores them by category in the database, and realizes query screening in the comprehensive query system; The related intelligence association multiplexing is based on the results of the video target structured analysis, and is associated and matched with other warning intelligence triggered in the same period. The access address of the forensic result is added to the warning intelligence with a higher correlation relationship to achieve the reuse of the forensic result.

Images