Perimeter safety prevention and control method and system based on light-vision linkage

Abstract

The invention belongs to the technical field of perimeter safety prevention and control, and discloses a perimeter safety prevention and control method and system based on optical-visual linkage, and the method comprises the steps: carrying out the target alarm detection of an intrusion event based on the synergistic effect of a blind spot error correction algorithm and a vibration ripple recognition engine in an optical fiber sensing technology, and carrying out the restoration of a vibration event; dynamically updating the blind spot error correction algorithm by using an incremental reinforcement learning mechanism; performing second-level intelligent video analysis and confirmation on the occurrence position of the vibration event by using a light-vision linkage calibration technology; performing intelligent perimeter detection and recognition alarm through a video intelligent analysis technology, and introducing a dynamic mask mechanism to suppress background noise; based on an event comprehensive judgment engine, performing comprehensive analysis on the alarm information to obtain a fusion alarm result; and executing a corresponding linkage action according to the fused alarm result. According to the invention, all-weather and high-sensitivity monitoring of the perimeter area is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of perimeter security control, and in particular to a perimeter security control method and system based on optical-visual linkage. Background Art

[0002] The importance of perimeter security for the South-to-North Water Diversion Project is self-evident. However, traditional perimeter protection methods are limited by monitoring range, warning accuracy, and susceptibility to external environmental interference (such as wind, rain, and obstruction), making them unable to meet the project's high perimeter security requirements. Currently, single perimeter detection technologies are unable to meet the security requirements of all-weather, zero missed alerts, and low false alarms.

[0003] The South-to-North Water Diversion Project's Middle Route stretches 1,432 kilometers, traversing diverse and complex terrains, including urban areas, rural areas, mountains, and farmland. The perimeter security system faces daunting challenges, including a wide range of terrain, numerous environmental interferences, and complex intrusion types. Traditional single technologies, such as infrared beamforming, vibrating fiber optics, and video surveillance, have significant shortcomings in practical applications. Infrared systems are susceptible to rain and fog, causing a sharp drop in sensitivity. Vibrating fiber optics have insufficient recognition for non-destructive intrusions (such as climbing and digging). Video surveillance has blind spots at night in low illumination and obscured by vegetation, and suffers from a generally high false alarm rate.

[0004] Currently, the industry generally uses the method of associating the optical fiber vibration area with the camera preset position to achieve the association between the vibration position and the video area. This method will consume a lot of manpower and material resources in the early stage of project delivery. In addition, when the pan-tilt head rotates a certain number of times and produces deviations, it cannot automatically adjust, which adds a lot of work for subsequent maintenance.

[0005] In addition, fiber optic sensing is the core technology layer of the optical-visual linkage system of the South-to-North Water Diversion Project's Middle Route. Its core component, the enhanced oDSP (Optical Digital Signal Processing) module, improves signal processing capabilities in complex environments through the collaborative innovation of a powerful blind spot error correction algorithm and a vibration ripple recognition engine. Although existing technologies are relatively mature, the following limitations have been identified in actual deployments:

[0006] Static model defects of blind spot error correction algorithms: Existing algorithms rely on offline trained error models and cannot adapt to progressive hardware degradation problems such as fiber aging and loose connectors; Insufficient generalization capabilities of vibration features: Existing engines have a high misjudgment rate for new intrusion modes (such as low-frequency mechanical resonance and multi-target superposition interference).

[0007] Therefore, how to provide a perimeter security control method and system based on optical and visual linkage is an urgent problem to be solved. Summary of the Invention

[0008] The embodiments of the present invention provide a perimeter security control method and system based on optical-visual linkage to solve the problems in the existing technology such as generally high false alarm rate, large subsequent maintenance workload, inability of existing blind spot error correction algorithms to adapt to progressive hardware degradation such as optical fiber aging and loose connectors, and insufficient generalization ability of vibration characteristics.

[0009] To provide a basic understanding of some aspects of the disclosed embodiments, the following is a brief summary. This summary is not intended to be a comprehensive review, identify key or essential elements, or delineate the scope of these embodiments. Its sole purpose is to present some concepts in a simplified form as a prelude to the detailed description that follows.

[0010] According to a first aspect of an embodiment of the present invention, a perimeter security control method based on optical-visual linkage is provided.

[0011] In one embodiment, a perimeter security control method based on optical and visual linkage includes:

[0012] Based on the synergy between the blind spot error correction algorithm in fiber optic sensing technology and the vibration ripple recognition engine, it can detect intrusion events with target alarms and restore vibration events. It also uses an incremental reinforcement learning mechanism to dynamically update the blind spot error correction algorithm.

[0013] Utilizing optical-visual linkage calibration technology, vibration events are located and confirmed through intelligent video analysis within seconds. Intelligent video analysis technology enables intelligent perimeter detection and identification, as well as the introduction of a dynamic masking mechanism to suppress background noise.

[0014] Based on the event comprehensive judgment engine, and according to the fiber optic sensing confidence, video sensing confidence, scene type and alarm discrimination strategy, the alarm information is comprehensively analyzed to obtain the fused alarm result; according to the fused alarm result, the corresponding linkage action is executed.

[0015] In one embodiment, based on the synergy between the blind spot error correction algorithm in the fiber optic sensing technology and the vibration ripple recognition engine, target alarm detection of intrusion events and restoration of vibration events include:

[0016] Detect intrusion events through sensing optical fibers to complete optical fiber sensing target alarm detection; send the detected optical fiber sensing target alarm information to the data access processing platform;

[0017] Using the blind spot error correction algorithm in fiber optic sensing technology, the signal phase collected by the intrusion event is corrected and shaped;

[0018] Utilizing the vibration ripple recognition engine in fiber optic sensing technology, vibration events are restored through multi-dimensional analysis to improve the accuracy of target alarm detection of intrusion events.

[0019] In one embodiment, dynamically updating the blind spot error correction algorithm using an incremental reinforcement learning mechanism includes:

[0020] An online learning module is embedded in the blind spot error correction algorithm to collect fiber link status data in real time and use it as environmental status input. The phase compensation parameter update strategy is dynamically optimized using the reduction in false alarm rate as the reward function. An elastic weight solidification algorithm is used to shorten the update cycle of the blind spot error correction model to hours.

[0021] In one embodiment, an online learning module is embedded in the blind spot error correction algorithm to collect fiber link status data in real time and use it as environmental status input. The phase compensation parameter update strategy is dynamically optimized using the reduction in false alarm rate as a reward function. An elastic weight solidification algorithm is used to shorten the blind spot error correction model update cycle to the hourly level. The following steps are included:

[0022] Construct a fiber link state space model and use the loss value and polarization mode dispersion coefficient as multi-dimensional state vector input;

[0023] Introducing anomaly detection algorithms in the dynamic environment perception layer to automatically identify link state mutation events and trigger reinforcement learning model updates.

[0024] The reduction in false alarm rate is used as the core indicator of the reward function, and a compound reward mechanism is constructed by combining immediate and delayed rewards.

[0025] Construct a dual-channel policy network, with the main network responsible for generating the phase compensation parameter adjustment strategy and the auxiliary network responsible for evaluating the long-term benefits of the strategy; and iteratively optimize the action selection mechanism through the Q function;

[0026] Use an experience replay pool to store historical state-action pairs and ensure a balance between knowledge inheritance and innovation;

[0027] Introducing the elastic weight curing algorithm to calculate the parameter importance matrix and constrain weight changes during the gradient descent process;

[0028] Establish an online verification sandbox to pre-execute newly generated policies in a virtual environment; once the evaluation meets the standards, deploy them to the physical layer;

[0029] Build a time-level model update trigger to force the model retraining process when the cumulative reward fluctuation exceeds the threshold.

[0030] In one embodiment, using optical-visual linkage calibration technology, performing second-level intelligent video analysis to confirm the location of a vibration event includes:

[0031] Through optical-visual linkage calibration technology, the associated camera is rotated to the location where the vibration event occurs, and the dome camera's perimeter intrusion intelligent recognition function is activated to perform video detection;

[0032] The servo autofocus method is used, and the target position is predicted through a prediction algorithm. The focus is dynamically adjusted to synchronize with the pan / tilt rotation speed to ensure that the image clarity meets the requirements when the pan / tilt rotates to the predetermined position, realizing intelligent recognition of the target object.

[0033] Among them, the illumination characteristics of the associated camera are used to locate objects at night, and the associated camera is driven to capture and track.

[0034] In one embodiment, intelligent perimeter detection and alarm recognition are performed using video intelligent analysis technology, and a dynamic mask mechanism is introduced to suppress background noise, including:

[0035] An intelligent perimeter detection algorithm is used to detect and identify targets. A dynamic masking mechanism is introduced based on actual project conditions to adaptively shield interference areas and suppress background noise.

[0036] In one embodiment, an intelligent perimeter detection algorithm is used to detect and identify the target. A dynamic mask mechanism is introduced based on the actual engineering situation to adaptively mask the interference area and suppress background noise.

[0037] Obtain interference area parameters and perform vector analysis. Through the inter-frame prediction mechanism, combined with the characteristics of motion vectors in different interference areas, a dynamic noise model is constructed. The optical flow motion result features are extracted to generate a dynamic weight mask.

[0038] Corresponding compensation algorithms and strategies are used for different interference areas to isolate background noise, and when the cumulative background difference exceeds the threshold, the full scene background is reconstructed;

[0039] Dynamically detect the target edge blur and adjust the threshold and compensation strategy; by adjusting the weight distribution, reduce the interference of the noise area while retaining the valid area.

[0040] In one embodiment, based on the event comprehensive decision engine, and according to the fiber optic sensing confidence, video sensing confidence, scene type, and alarm discrimination strategy, the alarm information is comprehensively analyzed to obtain the fused alarm results including:

[0041] Start the event comprehensive judgment engine and configure the fiber optic sensing confidence level and video sensing confidence level ranges to high, medium, and low;

[0042] Develop alarm discrimination strategies for different scenario types based on the confidence levels of fiber optic sensing and video sensing;

[0043] Based on the alarm discrimination strategy, the comprehensive discrimination confidence is output; the comprehensive discrimination confidence is used as the confidence of the fused alarm result.

[0044] In one embodiment, executing corresponding linkage actions according to the fusion alarm result includes:

[0045] The event access processing platform transmits the integrated alarm results to the security prevention management platform;

[0046] The security prevention management platform formulates different linkage actions based on the confidence level of the integrated alarm results;

[0047] Among them, the linkage actions include real-time video review, sound and light warning, expulsion and on-site disposal by patrol personnel.

[0048] According to a second aspect of an embodiment of the present invention, a perimeter security control system based on optical and visual linkage is provided.

[0049] In one embodiment, the perimeter security control system based on optical and visual linkage includes:

[0050] The fiber optic vibration detection module is used to detect intrusion targets and restore vibration events based on the synergy between the blind spot error correction algorithm in fiber optic sensing technology and the vibration ripple recognition engine. The blind spot error correction algorithm is dynamically updated using an incremental reinforcement learning mechanism.

[0051] The linkage detection module uses optical-visual linkage calibration technology to conduct intelligent video analysis and confirmation of the location of vibration events within seconds. It also uses intelligent video analysis technology to perform intelligent perimeter detection and identification alarms, and introduces a dynamic masking mechanism to suppress background noise.

[0052] The event comprehensive judgment and action execution module is used to conduct a comprehensive analysis of alarm information based on the event comprehensive judgment engine and according to the fiber optic sensor confidence, video sensor confidence, scene type and alarm discrimination strategy to obtain a fused alarm result; according to the fused alarm result, the corresponding linkage action is executed.

[0053] According to a third aspect of an embodiment of the present invention, a computer device is provided.

[0054] In some embodiments, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0055] According to a fourth aspect of embodiments of the present invention, a computer-readable storage medium is provided.

[0056] In one embodiment, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0057] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0058] (1) The present invention constructs a "light-visual linkage" multi-dimensional perception fusion technology system, which innovatively integrates distributed fiber optic sensing, intelligent video fixed-point analysis, weather (sunny, rainy, foggy, hazy, etc.), season (spring, summer, autumn, winter) and other multi-source heterogeneous data for deep integration. Among them, intelligent detection technology for fiber optic sensor echoes improves the positioning accuracy to ±2 meters; intelligent video uses deep learning algorithms to classify and identify intrusion behaviors and intrusion objects, and combines seasonal and meteorological factors to determine the proportion of alarms to build three-dimensional situational awareness; optical-visual linkage technology, as a combination of fiber optic sensing and intelligent vision technology, gives perimeter protection the combined advantages of multi-dimensional perception, multi-dimensional review and precise positioning. Through multi-modal data spatiotemporal alignment and evidence chain verification, the false alarm rate can be reduced to below 0.5% while ensuring zero missed reports. The effective warning rate in severe weather such as heavy rain and strong winds is increased to more than 98%, building a "zero missed report, low false alarm, all-weather, full coverage" protection and detection capability for complex perimeter scenarios, which serves as the preferred technical route for intelligent protection to build comprehensive security for the perimeter of the middle route of the South-to-North Water Diversion Project.

[0059] (2) The implementation of the present invention will significantly improve the efficiency and accuracy of perimeter protection for the South-to-North Water Diversion Project, and achieve rapid response and accurate handling of intrusion events through the functions of multi-dimensional perception, multi-dimensional verification, and precise positioning. At the same time, the technical solution of the present invention is highly integrated and complementary, and can give full play to the advantages of vibration fiber optic sensing technology and intelligent video monitoring technology, providing a solid guarantee for the security protection of the South-to-North Water Diversion Project. In addition, the method is also applicable to other large-scale, high-level perimeter applications, such as airports, railways, and pipeline networks, and has broad promotion value and application prospects.

[0060] (3) The present invention introduces an incremental reinforcement learning mechanism, embeds an online learning module in the blind spot error correction algorithm, collects optical fiber link status data (loss value, polarization mode dispersion coefficient) in real time as environmental state input, uses the reduction in false alarm rate as the reward function, and dynamically optimizes the phase compensation parameter update strategy; adopts an elastic weight solidification algorithm to prevent catastrophic forgetting of the model, and shortens the error correction model update cycle to hours.

[0061] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0063] Figure 1 This is a flow chart showing a perimeter security control method based on optical and visual linkage according to an exemplary embodiment;

[0064] Figure 2 This is a principle block diagram of a perimeter security control system based on optical and visual linkage according to an exemplary embodiment;

[0065] Figure 3 is a structural diagram of a computer device according to an exemplary embodiment;

[0066] Figure 4 is a diagram showing a deployment architecture according to an exemplary embodiment;

[0067] Figure 5 is a schematic diagram showing execution steps according to an exemplary embodiment;

[0068] Figure 6 The figure is a schematic diagram showing the interpretation rules of an event comprehensive decision engine according to an exemplary embodiment. DETAILED DESCRIPTION

[0069] The following description and accompanying drawings sufficiently illustrate the specific embodiments herein to enable those skilled in the art to practice them. Portions and features of some embodiments may be included in or substituted for portions and features of other embodiments. The scope of the embodiments herein includes the entire scope of the claims, including all available equivalents thereof. Herein, the terms "first," "second," and the like are used solely to distinguish one element from another and do not require or imply any actual relationship or order between these elements. In practice, the first element can also be referred to as the second element, and vice versa. Furthermore, the terms "comprise," "comprising," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a structure, device, or apparatus comprising a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such structure, device, or apparatus. Without further limitation, an element defined by the phrase "comprising a..." does not preclude the presence of other identical elements in the structure, device, or apparatus comprising the element. The various embodiments herein are described in a progressive manner, with each embodiment focusing on its differences from the other embodiments. Similar or identical parts between the various embodiments can be referenced to each other.

[0070] The terms "longitudinal", "transverse", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like used herein to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are intended only to facilitate the description of this document and simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the present invention. In the description herein, unless otherwise specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, they can be mechanical or electrical connections, or they can be internal connections between two elements, they can be directly connected, or they can be indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to the specific circumstances.

[0071] As used herein, unless otherwise specified, the term "plurality" means two or more.

[0072] In this document, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B means: A or B.

[0073] In this article, the term "and / or" is used to describe the association relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or, A and B.

[0074] It should be understood that, although the various steps in the flowchart are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps may be performed in other orders. Moreover, at least a portion of the steps in the figure may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily performed at the same time, but may be performed at different times. The execution order of these sub-steps or stages is not necessarily to be performed in sequence, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0075] Each module in the device or system of the present application can be implemented in whole or in part by software, hardware, or a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software so that the processor can call and execute the operations corresponding to the above modules.

[0076] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0077] Figure 1An embodiment of the perimeter security control method based on optical and visual linkage of the present invention is shown.

[0078] In this optional embodiment, the perimeter security control method based on optical and visual linkage includes:

[0079] S101. Based on the synergy between the blind spot error correction algorithm in fiber optic sensing technology and the vibration ripple recognition engine, target alarm detection is performed on intrusion events and vibration events are restored; the blind spot error correction algorithm is dynamically updated using the incremental reinforcement learning mechanism.

[0080] S102. Utilize optical-visual linkage calibration technology to conduct second-level intelligent video analysis and confirmation of the location of vibration events. Through intelligent video analysis technology, intelligent perimeter detection and identification alarms are performed, and a dynamic mask mechanism is introduced to suppress background noise.

[0081] S103. Based on the event comprehensive judgment engine, and according to the fiber optic sensing confidence, video sensing confidence, scene type and alarm discrimination strategy, the alarm information is comprehensively analyzed to obtain a fused alarm result; according to the fused alarm result, the corresponding linkage action is executed.

[0082] In this optional embodiment, based on the synergy between the blind spot error correction algorithm in the optical fiber sensing technology and the vibration ripple recognition engine, target alarm detection of intrusion events and restoration of vibration events include:

[0083] Intrusion events are detected through sensing optical fibers to complete optical fiber sensing target alarm detection; the detected optical fiber sensing target alarm information is sent to the data access and processing platform; the blind spot error correction algorithm in optical fiber sensing technology is used to correct and shape the signal phase collected by the intrusion event; the vibration ripple recognition engine in optical fiber sensing technology is used to restore vibration events through multi-dimensional analysis to improve the accuracy of target alarm detection of intrusion events.

[0084] In this optional embodiment, dynamically updating the blind spot error correction algorithm using an incremental reinforcement learning mechanism includes:

[0085] An online learning module is embedded in the blind spot error correction algorithm to collect fiber link status data in real time and use it as environmental status input. The phase compensation parameter update strategy is dynamically optimized using the reduction in false alarm rate as the reward function. An elastic weight solidification algorithm is used to shorten the update cycle of the blind spot error correction model to hours.

[0086] In this optional embodiment, an online learning module is embedded in the blind spot error correction algorithm to collect fiber link status data in real time and use it as environmental status input. The phase compensation parameter update strategy is dynamically optimized using the reduction in false alarm rate as a reward function. An elastic weight solidification algorithm is used to shorten the blind spot error correction model update cycle to the hourly level. The following are included:

[0087] A fiber optic link state space model is constructed, and the loss value and polarization mode dispersion coefficient are used as multi-dimensional state vector inputs; an anomaly detection algorithm is introduced in the dynamic environment perception layer to automatically identify link state mutation events and trigger the reinforcement learning model update response; the reduction in false alarm rate is used as the core indicator of the reward function, and a composite reward mechanism is constructed by combining immediate rewards and delayed rewards; a dual-channel strategy network is constructed, and the main network is responsible for generating phase compensation parameter adjustment strategies to assist the network in evaluating the long-term benefits of the strategy; the action selection mechanism is optimized through Q function iteration; an experience replay pool is used to store historical state-action pairs and ensure a balance between knowledge inheritance and innovation; an elastic weight solidification algorithm is introduced to calculate the parameter importance matrix and constrain weight changes during the gradient descent process; an online verification sandbox is established to pre-execute the newly generated strategy in a virtual environment; and it is deployed to the physical layer when the evaluation meets the standards; a time-level model update trigger is constructed to force the model retraining process to start when the cumulative reward fluctuation exceeds the threshold.

[0088] In this optional embodiment, using optical-visual linkage calibration technology to perform second-level intelligent video analysis and confirmation of the location of the vibration event includes:

[0089] Through optical-visual linkage calibration technology, the associated camera is rotated to the location where the vibration event occurs, and the perimeter intrusion intelligent recognition function of the ball camera is activated for video detection; the servo autofocus method is adopted, and the target position is predicted through the prediction algorithm, and the focal length is dynamically adjusted to synchronize with the pan-tilt head steering speed to ensure that the image clarity meets the requirements when the pan-tilt head rotates to the predetermined position, thereby realizing intelligent recognition of the target object; among them, the illumination characteristics of the associated camera are used to realize the positioning of objects at night, and drive the associated camera to capture and track.

[0090] In this optional embodiment, intelligent perimeter detection and recognition alarms are performed through video intelligent analysis technology, and a dynamic mask mechanism is introduced to suppress background noise, including:

[0091] An intelligent perimeter detection algorithm is used to detect and identify targets. A dynamic masking mechanism is introduced based on actual project conditions to adaptively shield interference areas and suppress background noise.

[0092] In this optional embodiment, the intelligent perimeter detection algorithm is used to detect and identify the target, and a dynamic mask mechanism is introduced in combination with the actual engineering situation to adaptively mask the interference area and suppress background noise.

[0093] Obtain interference area parameters and perform vector analysis. Through the inter-frame prediction mechanism, combined with the characteristics of motion vector performance in different interference areas, a dynamic noise model is constructed; the optical flow motion result features are extracted to generate a dynamic weight mask; corresponding compensation algorithms and strategies are used for different interference areas to isolate background noise, and when the cumulative background difference exceeds the threshold, the full scene background is reconstructed; the target edge blur is dynamically detected, and the threshold and compensation strategy are adjusted; by adjusting the weight distribution, the interference of the noise area is reduced while retaining the effective area.

[0094] In this optional embodiment, based on the event comprehensive decision engine, and according to the optical fiber sensing confidence, video sensing confidence, scene type and alarm discrimination strategy, the alarm information is comprehensively analyzed to obtain the fused alarm results including:

[0095] Start the event comprehensive judgment engine and configure the range of fiber optic sensing confidence and video sensing confidence to high, medium and low; formulate alarm judgment strategies for different scenario types based on the fiber optic sensing confidence and video sensing confidence; output the comprehensive judgment confidence based on the alarm judgment strategy; use the comprehensive judgment confidence as the confidence of the fused alarm result.

[0096] In this optional embodiment, executing corresponding linkage actions according to the fusion alarm result includes:

[0097] The event access and processing platform transmits the integrated alarm results to the security prevention management platform; the security prevention management platform formulates different linkage actions based on the confidence level of the integrated alarm results; among them, linkage actions include real-time video review, sound and light warning, expulsion and on-site disposal by patrol personnel.

[0098] Figure 2 An embodiment of the perimeter security control system based on optical and visual linkage of the present invention is shown.

[0099] In this optional embodiment, the perimeter security control system based on optical and visual linkage includes:

[0100] The optical fiber vibration detection module 201 is used to perform target alarm detection of intrusion events and restore vibration events based on the synergy between the blind spot error correction algorithm and the vibration ripple recognition engine in the optical fiber sensing technology; and dynamically update the blind spot error correction algorithm using the incremental reinforcement learning mechanism.

[0101] The linkage detection module 202 is used to use optical-visual linkage calibration technology to perform second-level intelligent video analysis to confirm the location of vibration events; through video intelligent analysis technology, it performs intelligent perimeter detection and identification alarms, and introduces a dynamic mask mechanism to suppress background noise.

[0102] The event comprehensive judgment and action execution module 203 is used to comprehensively analyze the alarm information based on the event comprehensive judgment engine and according to the fiber optic sensing confidence, video sensing confidence, scene type and alarm discrimination strategy to obtain a fused alarm result; and execute corresponding linkage actions according to the fused alarm result.

[0103] In order to facilitate understanding of the above technical solutions of the present invention, the above technical solutions of the present invention are further explained from the perspective of architecture and principle as follows:

[0104] This invention combines novel fiber-optic vibration sensing technology with the optical-visual linkage of intelligent video surveillance. Designed specifically for complex and demanding perimeter scenarios like the South-to-North Water Diversion Project, it provides an efficient, reliable, and comprehensive security monitoring solution. This invention explores the integration of multiple sensing technologies to leverage their strengths and address their weaknesses, better meeting the requirements for perimeter protection of the South-to-North Water Diversion Project's central route. By combining the advantages of vibration fiber-optic sensing and intelligent video surveillance, it achieves all-weather, high-sensitivity monitoring of the perimeter area and provides detailed and intuitive visual information, significantly improving the efficiency and accuracy of perimeter protection.

[0105] The invention has the following key technologies:

[0106] 1. Optical and visual joint calibration technology

[0107] The present invention relies on longitude, latitude, relative altitude and camera pitch angle, calculates the relative position of the optical fiber vibration point and the camera, and achieves rapid video positioning through precise control of the pan-tilt head. It improves delivery and operation and maintenance efficiency and enhances positioning accuracy through a small amount of calculation.

[0108] 2. Optical and visual fast linkage technology

[0109] In the perimeter protection scenario of the South-to-North Water Diversion Project's central route, to facilitate rapid response, there are clear timeliness requirements for detecting target intrusions, with alarm detection required to be completed within 5 seconds. However, the optical-visual linkage processing involves multiple links and requires an end-to-end design to ensure that latency performance targets are achieved. This invention improves the efficiency of optical-visual linkage through the following methods:

[0110] a) Reduced fiber-optic sensing triggering latency: Traditional fiber-optic sensing triggering of video linkage has a large latency. Using short recognition can trigger camera linkage in advance, achieving dual sensing in parallel and significantly improving optical-visual linkage performance.

[0111] b) Fast camera focus: The camera pan / tilt rotates to focus the image, which may take seconds at night. The present invention supports fast focus, that is, focusing is completed during the rotation process, saving focusing time and shortening video perception delay.

[0112] c) Real-time task scheduling: The management and control platform introduces real-time task scheduling to increase the priority of linkage tasks, ensure that linkage control and perception results are completed in sub-seconds, and ensure stable optical-visual linkage performance.

[0113] 3. Comprehensive judgment of the incident

[0114] Traditional optical-visual linkage uses a simple AND logic between fiber and video. An intrusion alarm is only reported if both fiber and video sensors detect an alarm simultaneously. This mechanism significantly increases the risk of missed alarms in the optical-visual linkage. The present invention's comprehensive judgment makes a comprehensive assessment based on the confidence levels of both fiber and video sensing. When video sensing is unreliable, fiber sensing is primarily relied upon, and vice versa. This approach leverages the respective strengths of fiber and video, while simultaneously addressing both missed and false alarms, achieving the optimal judgment result.

[0115] The technical solution of the present invention includes the following core parts:

[0116] 1. Multi-dimensional perception

[0117] Deploy a vibration fiber-optic sensor network to detect vibration signals in the perimeter in real time, ensuring a rapid response to any intrusion. Intelligent analysis of fiber-optic vibration signals ensures accurate analysis, reduces false alarm rates, and improves accuracy. Utilize intelligent video surveillance technology to provide comprehensive, all-encompassing video surveillance of the perimeter where vibration alarms occur.

[0118] 2. Multi-dimensional review mechanism

[0119] Vibration fiber optic sensing technology is tightly integrated with intelligent video surveillance technology to provide multi-dimensional and multi-level verification of detected intrusion events. Combined with intelligent video analysis technology, intelligent analysis of the area where the vibration alarm occurred and intelligent detection of related areas enable secondary intelligent verification of optical vibration alarms, providing intelligent visual information to supplement and verify the alarm, improving alarm accuracy and reducing false alarm rates. Local weather conditions are also used to determine the weighting of optical vibration and video intelligent analysis results. In good weather, the video intelligent analysis results are given a higher weighting; otherwise, the optical fiber vibration results are given a higher weighting. Subsequent alarm results are evaluated based on the combined results of these two factors, resulting in a final conclusion to ensure the accuracy of the alarm information.

[0120] 3. Accurate positioning function

[0121] Leveraging the precise positioning capabilities of vibrating fiber optic sensing technology, the intruder's location can be quickly determined, providing instant information for emergency response. Combined with the image recognition capabilities of video surveillance technology, the intruder can be further precisely identified, providing conclusive evidence for subsequent action.

[0122] 4. High integration and complementarity

[0123] The high level of integration of vibration fiber optic sensing technology and intelligent video surveillance technology enables complementary advantages and synergistic operation. This integration not only improves the overall performance of the system, but also reduces operation and maintenance costs, improving the cost-effectiveness of perimeter protection.

[0124] This invention proposes a perimeter security control method for the South-to-North Water Diversion Project based on optical-visual linkage. This method combines vibrating optical fibers with intelligent video analysis to achieve multi-dimensional perimeter security perception and control, improving accuracy and reducing false alarm and missed alarm rates. To achieve this objective, the invention includes the following implementation steps (preliminary completion of optical-visual linkage configuration):

[0125] The deployment architecture is divided into three layers:

[0126] 1. Optical-visual linkage deployment on the edge: Optical sensing devices, including fiber optic sensors and video sensors, are deployed on the edge to collect perimeter intrusion signals. Optical sensing uses vibrating fiber optics, while video sensing uses dome cameras or box cameras, with dome cameras being the primary choice.

[0127] 2. Optical-visual integration edge deployment: Fiber-optic sensing and analysis equipment and video access storage devices are deployed at the edge to analyze, detect, and forward single-sensing information. Fiber-optic sensing devices primarily transmit and receive fiber-optic signals, using changes in those signals to detect perimeter intrusions. Video access storage devices receive, store, and access videos. They control cameras based on the control platform, enabling pan / tilt (PTZ) rotation and image capture. They also receive intelligent perimeter intrusion alerts from cameras and distribute them to the control platform.

[0128] 3. Deployment at the optical-visual linkage center: The data access and processing platform supports correlation mapping between the two types of sensing, enabling the coordinated operation of fiber-optic and video sensing. It also performs correlation, fusion, and collision analysis on the fiber-optic and video sensing results, outputting the final optical-visual linkage detection results. The security management platform connects to the perimeter alarm data from the data access and processing platform and provides coordinated alarm prompts based on the security requirements of the South-to-North Water Diversion Project's middle route.

[0129] The specific architecture is as follows Figure 4 Specific execution steps are as follows: Figure 5 . Figure 4 The center side includes a security management platform and a data access and processing platform; the edge side includes fiber optic sensing, which is used to transmit light alarms to the data access and processing platform, and also includes video access / storage, which is used to transmit camera intelligent alarm images / videos to the data access and processing platform. Figure 5It includes: the sensor fiber performs fiber optic sensing and sends the fiber optic sensing target to the data access processing platform; the data access processing platform notifies the video linkage detection to the video access / storage, and transmits the camera linkage detection to the video sensor (smart dome camera); the video sensor (smart dome camera) transmits the camera detection result to the video access / storage, and the video access / storage reports the video alarm to the data access processing platform; the data access processing platform makes a comprehensive judgment, and the data access processing platform and the security prevention management platform realize the notification business platform.

[0130] Step 1: Real-time detection of optical fiber vibration

[0131] When an intrusion event occurs, it is detected via the sensing fiber, identified through fiber-optic sensing, and alerted to the target. The alert information is then sent to the data access and processing platform. The enhanced oDSP (Optical Digital Signal Processing) module in the fiber-optic sensing section incorporates a robust blind-spot error correction algorithm, which corrects and reshapes the phase of the captured weak signal, significantly improving its effectiveness. A built-in vibration ripple recognition engine also enables multi-dimensional analysis and restoration of vibration events. For each vibration point, at least 32 phases of information are captured, and multi-dimensional features (such as voiceprint, frequency, spatial information, timing, and duration) are extracted. Multi-dimensional deep convolution is then used to identify and compare samples, significantly improving the accuracy of the analysis results.

[0132] To address the shortcomings of the existing technology, the present invention introduces an incremental reinforcement learning mechanism, embeds an online learning module in the blind spot error correction algorithm, collects optical fiber link status data (loss value, polarization mode dispersion coefficient) in real time as environmental state input, uses the reduction in false alarm rate as the reward function, and dynamically optimizes the phase compensation parameter update strategy; adopts an elastic weight solidification algorithm to prevent catastrophic forgetting of the model, shortening the error correction model update cycle to hours.

[0133] The specific implementation steps are as follows:

[0134] 1. Environmental state modeling and real-time collection

[0135] A fiber link state space model is constructed, using loss and polarization mode dispersion coefficient as multidimensional state vector inputs. Online monitoring is used to collect raw data at a minute-by-minute frequency, and a sliding window mechanism is used to extract time series features. The dynamic environment perception layer incorporates an anomaly detection algorithm to automatically identify sudden link state events and trigger updates to the reinforcement learning model.

[0136] 2. Reward Function and Policy Network Design

[0137] The reduction in false alarm rate is defined as the core metric of the reward function. A composite reward mechanism is constructed by combining immediate rewards (for the effectiveness of a single error correction) with delayed rewards (for sustained stability). A dual-channel policy network is designed: the primary network generates the phase compensation parameter adjustment strategy, while the auxiliary network evaluates the long-term benefits of the strategy. The action selection mechanism is iteratively optimized using the Q function, which is used to estimate the expected long-term cumulative reward.

[0138] 3. Incremental online learning

[0139] An experience replay pool is used to store historical state-action pairs. New and old samples are mixed in a 7:3 ratio during each model update to ensure a balance between knowledge inheritance and innovation. An elastic weight curation algorithm is introduced to calculate a parameter importance matrix, constraining changes in key weights during gradient descent to prevent catastrophic forgetting.

[0140] 4. Dynamic parameter optimization and verification

[0141] Establish an online verification sandbox to pre-execute newly generated policies in a virtual environment and deploy them to the physical layer after evaluation and compliance.

[0142] Design hourly model update triggers: When the cumulative reward fluctuation exceeds the threshold, the model retraining process is forced to start.

[0143] Step 2: Video linkage detection

[0144] The data access and processing platform controls the event-related camera, which uses optical-visual linkage calibration technology to rotate to the location of the vibration event. Simultaneously, the dome camera's intelligent perimeter intrusion recognition function is activated for video detection. This invention abandons traditional preset calibration schemes and adopts a method based on relative longitude and latitude and elevation differences to achieve stepless and rapid positioning of the dome camera, reducing system deployment intensity and improving deployment efficiency.

[0145] Based on the relative longitude and latitude and height difference, the speed dome camera can achieve stepless and rapid positioning. The principle of implementation is: using the servo autofocus method, predicting the target position through prediction algorithms (such as Kalman filtering), dynamically adjusting the focal length and PTZ steering speed to ensure the clarity of the image when the PTZ rotates to the predetermined position, and realizing rapid and intelligent recognition of the target object.

[0146] To ensure real-time analysis, the camera achieves rapid focusing during its rotation, leveraging its low-light capabilities to quickly locate objects at night. This enables the camera's intelligent capture and tracking functions after analysis. This strategy enables intelligent video analysis and confirmation of fiber optic vibration within seconds.

[0147] Step 3: Video Intelligent Analysis

[0148] After driving the camera to the specified position and angle, the intelligent perimeter detection algorithm built into the smart camera is started to perform target detection and recognition, and the detection and recognition results are output within a given time. In this invention, the intelligent perimeter detection algorithm is embedded with the help of the built-in computing power of the camera. Combined with the actual situation of the South-to-North Water Diversion Project, a dynamic mask mechanism is introduced on the basis of the original general algorithm to adaptively shield interference areas such as vegetation shaking and water surface reflection, suppress background noise and improve target detection efficiency.

[0149] Specific process and steps:

[0150] 1. Interference area detection and modeling

[0151] Polarized imaging is used to capture water surface reflections and vegetation sway parameters. Motion vector analysis is performed on these two types of areas. Using an inter-frame prediction mechanism, a dynamic noise model is constructed based on the characteristics of motion vectors in water surface reflections and vegetation sway. The model is then used to extract optical flow motion features and generate a dynamic weighted mask.

[0152] 2. Noise suppression and target enhancement

[0153] Different compensation algorithms and strategies (polarization compensation for water and motion compensation for vegetation) are used to isolate background noise in reflective areas of the water surface and areas with swaying vegetation. When the cumulative background difference exceeds a threshold (e.g., SSIM < 0.8), the full scene background is reconstructed.

[0154] 3. Real-time optimization and performance tuning

[0155] Online parameter tuning: Dynamically detects target edge blur, adjusts thresholds and compensation strategies. By adjusting weight distribution, it reduces interference in noisy areas while retaining the effective areas, thereby improving target detection efficiency.

[0156] The detection results (including pictures, target categories, etc.) are transmitted to the data access and processing platform through the video access / storage module.

[0157] Step 4: Comprehensive judgment of the incident

[0158] The event access processing platform has a built-in event comprehensive judgment engine, which conducts comprehensive analysis on the received optical fiber vibration alarm and video intelligent recognition alarm, and generates the final integrated alarm information according to the preset conditions (time, space, regular expression, etc.); the event comprehensive judgment engine interprets the rules such as Figure 6Among them, scene types include normal conditions with good lighting conditions and no obstruction; night, thick fog, haze, obstruction; wind, rain, snow, hail; large vehicles passing by; extreme wind, rain, snow, and hail weather; alarm discrimination strategies include both (fiber optic sensor confidence and video sensor confidence) with high confidence, comprehensive discrimination: high; discrimination strategy based on fiber optic sensing, discrimination confidence: high; discrimination strategy based on video sensing, discrimination confidence: high; discrimination strategy based on video sensing, discrimination confidence: high; both confidence levels are low, comprehensive discrimination: low. The above alarm discrimination strategies correspond to scene types.

[0159] Step 5: Business scenario application

[0160] The event access and processing platform will transmit the integrated alarm results to the security prevention management platform. The security prevention management platform will formulate different linkage actions based on the confidence level of the alarm, including but not limited to real-time video review, sound and light alarm, evacuation, and on-site disposal by patrol personnel.

[0161] This invention combines the advantages of vibration fiber optic sensing and intelligent video surveillance technologies to achieve multi-dimensional perception, multi-dimensional verification, and precise positioning. It also involves deploying a vibration fiber optic sensing network to sense vibration signals within the perimeter in real time, and then processing and analyzing these signals in real time to detect intrusions. Furthermore, it incorporates intelligent video surveillance technology to provide comprehensive, comprehensive video surveillance of the perimeter, providing detailed monitoring information and evidence collection capabilities to ensure accurate recording and tracing of intrusions.

[0162] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store static information and dynamic information data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the steps of the above-mentioned method embodiment are implemented.

[0163] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0164] In addition, the present invention also provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiment when executing the computer program.

[0165] In addition, the present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiment are implemented.

[0166] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0167] The present invention is not limited to the structures described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A perimeter security control method based on optical and visual linkage, characterized in that: include: Based on the synergy between the blind spot error correction algorithm in fiber optic sensing technology and the vibration ripple recognition engine, it can detect intrusions and restore vibration events. The blind spot error correction algorithm is dynamically updated using an incremental reinforcement learning mechanism. Utilizing optical-visual linkage calibration technology, vibration events are located through intelligent video analysis within seconds. Intelligent video analysis technology enables intelligent perimeter detection and identification, as well as the introduction of a dynamic masking mechanism to suppress background noise. Based on the event comprehensive judgment engine, and according to the fiber optic sensing confidence, video sensing confidence, scene type and alarm discrimination strategy, the alarm information is comprehensively analyzed to obtain the integrated alarm result; Execute corresponding linkage actions based on the fusion alarm results.

2. A perimeter security control method based on optical and visual linkage according to claim 1, characterized in that: The synergy between the blind spot error correction algorithm in the fiber optic sensing technology and the vibration ripple recognition engine to detect intrusion targets and restore vibration events includes: Detect intrusion events through sensing optical fibers to complete optical fiber sensing target alarm detection; send the detected optical fiber sensing target alarm information to the data access processing platform; Using the blind spot error correction algorithm in fiber optic sensing technology, the signal phase collected by the intrusion event is corrected and shaped; Utilizing the vibration ripple recognition engine in fiber optic sensing technology, vibration events are restored through multi-dimensional analysis to improve the accuracy of target alarm detection of intrusion events.

3. The perimeter security control method based on optical and visual linkage according to claim 1 is characterized in that: The method of dynamically updating the blind spot error correction algorithm by using the incremental reinforcement learning mechanism includes: An online learning module is embedded in the blind spot error correction algorithm to collect fiber link status data in real time and use it as environmental status input. The phase compensation parameter update strategy is dynamically optimized using the reduction in false alarm rate as the reward function. An elastic weight solidification algorithm is used to shorten the update cycle of the blind spot error correction model to hours.

4. The perimeter security control method based on optical and visual linkage according to claim 3 is characterized in that: The online learning module is embedded in the blind spot error correction algorithm to collect optical fiber link status data in real time and use it as environmental status input. The phase compensation parameter update strategy is dynamically optimized using the reduction in false alarm rate as a reward function. Adopting an elastic weight solidification algorithm shortens the blind spot error correction model update cycle to hours, including: Construct a fiber link state space model and use the loss value and polarization mode dispersion coefficient as multi-dimensional state vector input; Introducing anomaly detection algorithms in the dynamic environment perception layer to automatically identify link state mutation events and trigger reinforcement learning model updates. The reduction in false alarm rate is used as the core indicator of the reward function, and a compound reward mechanism is constructed by combining immediate and delayed rewards. Construct a dual-channel policy network, with the main network responsible for generating the phase compensation parameter adjustment strategy and the auxiliary network responsible for evaluating the long-term benefits of the strategy; and iteratively optimize the action selection mechanism through the Q function; Use an experience replay pool to store historical state-action pairs and ensure a balance between knowledge inheritance and innovation; Introducing the elastic weight curing algorithm to calculate the parameter importance matrix and constrain weight changes during the gradient descent process; Establish an online verification sandbox to pre-execute newly generated policies in a virtual environment; once the evaluation meets the standards, deploy them to the physical layer; Build a time-level model update trigger to force the model retraining process when the cumulative reward fluctuation exceeds the threshold.

5. The perimeter security control method based on optical and visual linkage according to claim 1 is characterized in that: The optical-visual linkage calibration technology is used to perform second-level intelligent video analysis and confirmation of the location of the vibration event, including: Through optical-visual linkage calibration technology, the associated camera is rotated to the location where the vibration event occurs, and the dome camera's perimeter intrusion intelligent recognition function is activated to perform video detection; The servo autofocus method is used, and the target position is predicted through a prediction algorithm. The focus is dynamically adjusted to synchronize with the pan / tilt rotation speed to ensure that the image clarity meets the requirements when the pan / tilt rotates to the predetermined position, realizing intelligent recognition of the target object. Among them, the illumination characteristics of the associated camera are used to locate objects at night, and the associated camera is driven to capture and track.

6. The perimeter security control method based on optical and visual linkage according to claim 1 is characterized in that: The aforementioned intelligent perimeter detection and alarm recognition using video intelligent analysis technology, and the introduction of a dynamic masking mechanism to suppress background noise, include: Utilize intelligent perimeter detection algorithms to detect and identify targets. Combined with actual project conditions, a dynamic masking mechanism is introduced to adaptively shield interference areas and suppress background noise.

7. The perimeter security control method based on optical and visual linkage according to claim 6, characterized in that: The intelligent perimeter detection algorithm is used to detect and identify targets. Based on the actual engineering situation, a dynamic masking mechanism is introduced to adaptively mask the interference area and suppress background noise. The following methods are included: Obtain interference area parameters and perform vector analysis. Through the inter-frame prediction mechanism, combined with the characteristics of motion vectors in different interference areas, a dynamic noise model is constructed. The optical flow motion result features are extracted to generate a dynamic weight mask. Corresponding compensation algorithms and strategies are used for different interference areas to isolate background noise, and when the cumulative background difference exceeds the threshold, the full scene background is reconstructed; Dynamically detect the target edge blur and adjust the threshold and compensation strategy; by adjusting the weight distribution, reduce the interference of the noise area while retaining the valid area.

8. The perimeter security control method based on optical and visual linkage according to claim 1 is characterized in that: The event-based comprehensive judgment engine performs a comprehensive analysis of alarm information based on the fiber optic sensing confidence, video sensing confidence, scene type, and alarm discrimination strategy, and obtains the following integrated alarm results: Start the event comprehensive judgment engine and configure the fiber optic sensing confidence level and video sensing confidence level ranges to high, medium, and low; Develop alarm discrimination strategies for different scenario types based on the confidence levels of fiber optic sensing and video sensing; Based on the alarm discrimination strategy, the comprehensive discrimination confidence is output; the comprehensive discrimination confidence is used as the confidence of the fused alarm result.

9. The perimeter security control method based on optical and visual linkage according to claim 1, characterized in that: The execution of corresponding linkage actions according to the fusion alarm result includes: The event access processing platform transmits the integrated alarm results to the security prevention management platform; The security prevention management platform formulates different linkage actions based on the confidence level of the integrated alarm results; Among them, the linkage actions include real-time video review, sound and light warning, expulsion and on-site disposal by patrol personnel.

10. A perimeter security control system based on optical and visual linkage, characterized in that: include: The fiber optic vibration detection module is used to detect intrusion targets and restore vibration events based on the synergy between the blind spot error correction algorithm in fiber optic sensing technology and the vibration ripple recognition engine. The blind spot error correction algorithm is dynamically updated using an incremental reinforcement learning mechanism. The linkage detection module uses optical-visual linkage calibration technology to perform intelligent video analysis and confirmation of the location of vibration events within seconds. It also uses intelligent video analysis technology to perform intelligent perimeter detection and identification alarms, and introduces a dynamic masking mechanism to suppress background noise. The event comprehensive judgment and action execution module is used to comprehensively analyze the alarm information based on the event comprehensive judgment engine and according to the fiber optic sensor confidence, video sensor confidence, scene type and alarm discrimination strategy to obtain the integrated alarm result; Execute corresponding linkage actions based on the fusion alarm results.

Images