Security and protection system equipment operation state self-diagnosis and predictive maintenance method

Abstract

The invention provides a security and protection system equipment operation state self-diagnosis and predictive maintenance method, which relates to the field of equipment monitoring and diagnosis control, and comprises the following steps of: acquiring cooperative motion data of each pan-tilt motor in a camera array, calculating a synchronization index of the motion of adjacent camera pan-tilts, and constructing a pan-tilt group control cooperative matrix; therefore, the cooperative motion characteristics of the holder equipment are comprehensively analyzed. On the basis, the load balance degree and the optimized overlapping area of each camera holder are calculated, and a dynamic task allocation strategy is adopted to carry out grouping control on a holder group, so that the optimized scheduling of a multi-target tracking task is effectively realized. According to the method, the cooperative work efficiency of the holder group can be remarkably improved, blind areas and invalid overlapping areas of a monitoring area can be reduced, and the overall monitoring effect is optimized.

Description

Technical Field

[0001] This invention relates to the field of equipment monitoring and diagnostic control, and more specifically, to a method for self-diagnosis and predictive maintenance of the operating status of security system equipment. Background Technology

[0002] With the development of modern security technology, video surveillance systems have become a core component of social security. In security applications, camera pan-tilt units (PTZ) are widely used for high-density target monitoring, intelligent behavior analysis, and multi-target tracking in complex environments due to their flexible rotation and precise positioning capabilities. However, with the increasing demand for intelligent security, the complexity and operational pressure of video surveillance systems have also significantly increased. For example, in a large-scale monitoring network, camera PTZ devices need to simultaneously handle real-time tracking of multiple targets while maintaining efficient and accurate collaborative operation, thus placing higher demands on their motion performance, load balancing, and long-term stability. To meet these requirements, researchers have developed many advanced PTZ control algorithms and device management technologies in recent years, such as AI-based target recognition and tracking technologies and dynamic programming-based resource allocation algorithms. However, existing technologies mainly focus on improving the intelligence level of individual camera PTZ units, lacking systematic research on the collaborative efficiency and long-term operational reliability of PTZ groups, and cannot effectively address the performance degradation caused by equipment aging or collaborative misalignment in complex monitoring scenarios.

[0003] Existing security systems often suffer from several shortcomings that negatively impact overall performance and maintenance efficiency. Firstly, in multi-target tracking tasks, the load distribution of camera pan-tilt units (PTZs) often lacks an overall optimization strategy, easily leading to some devices operating under overload for extended periods, accelerating wear and tear and reducing lifespan. Furthermore, motion trajectory optimization and task allocation in PTZ group control fail to adequately consider inter-device coordination and consistency, resulting in ineffective overlap or blind spots between camera monitoring areas, reducing monitoring effectiveness. Secondly, traditional equipment maintenance methods typically rely on periodic maintenance or post-failure repair, lacking real-time diagnosis and predictive analysis of PTZ device operating status. This passive maintenance strategy not only increases the risk of equipment downtime but also significantly raises system maintenance costs. Therefore, how to achieve collaborative optimization control of PTZ devices through scientific algorithms and realize self-diagnosis and predictive maintenance during equipment operation has become a critical issue urgently needing to be addressed in the current security technology field. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention is proposed. This invention provides a method for self-diagnosis and predictive maintenance of the operating status of security system equipment, which can, to some extent, solve the problems of pan-tilt-zoom (PTZ) group control conflicts and accelerated equipment wear caused by multi-target tracking in camera arrays.

[0005] According to one aspect of the present invention, a method for self-diagnosis and predictive maintenance of the operating status of security system equipment is provided, comprising: Collect the coordinated motion data of each gimbal motor in the camera array, and construct a gimbal group control coordination matrix by calculating the synchronization index of the motion of adjacent camera gimbals. Based on the aforementioned gimbal group control and coordination matrix, the load balance of each camera gimbal during multi-target tracking is calculated, and an optimized overlapping area of ​​the gimbal motion trajectory is established. Based on the load balancing degree and optimized overlapping area, a dynamic task allocation strategy is adopted to group and control the PTZ group, and the target tracking task of each PTZ is determined by calculating the switching loss between PTZ groups. Adaptive smoothing is applied to the motion trajectories of each gimbal group, and a gimbal wear assessment report is generated by analyzing the cooperative motion characteristics of the gimbal groups.

[0006] Furthermore, the calculation of the synchronicity index includes: The position matching degree is calculated based on the ratio of the gimbal position deviation to the reference distance; The horizontal and vertical velocity vectors are calculated based on the gimbal rotation speed data, and the speed coordination is obtained by the cosine value of the angle between the two velocity vectors. The load correlation is obtained by calculating the Pearson correlation coefficient of the normalized current waveforms of the two gimbals. The synchronization index is obtained by weighted summation of the position matching degree, the speed coordination degree, and the load correlation degree.

[0007] Furthermore, the load balancing degree is calculated by combining the collaborative load vector and dynamic load vector of the gimbal in a spherical coordinate system into a comprehensive load state vector. The cooperative load vector is obtained by superimposing vectors of non-zero elements in polar coordinates; The dynamic load vector is obtained by calculating the trajectory curvature change and velocity distribution characteristics of the gimbal during the most recent detection period.

[0008] Furthermore, the optimized overlapping region of the gimbal motion trajectory is established by calculating the divergence field of the comprehensive load state vector of adjacent gimbals to identify the load distribution characteristics. Combined with the circulation direction of the curl field, the dynamic optimization boundary of the gimbal motion trajectory is determined under the condition of satisfying the boundary continuity.

[0009] Furthermore, the load distribution characteristics are identified by performing local statistics on the divergence values ​​using a sliding block. Regions with a negative divergence mean and small variance are marked as load accumulation regions, while regions with a positive divergence mean and small variance are marked as load sparse regions.

[0010] Furthermore, the boundary continuity is determined based on a boundary growth algorithm; The boundary growth algorithm expands the boundary tangentially along the curl field within the dynamic equilibrium range based on the ratio of divergence to curl at the boundary point. When the absolute value of divergence increases significantly, the boundary orientation is adjusted along the divergence gradient direction.

[0011] Furthermore, the boundary growth algorithm is shown in the following equation: , in, , in, Let be the boundary point position vector at time t. For time step, The boundary growth rate vector. The tangential spreading rate coefficient, For normal adjustment rate coefficient, Let be the tangential unit vector of the curl field. Let be the divergence field scalar. The curl field vector, The divergence-curl ratio threshold, This is the divergence threshold.

[0012] Furthermore, the dynamic task allocation strategy optimizes the task allocation scheme by grouping gimbals according to load balancing and then based on the target density distribution of overlapping areas and the switching loss between gimbal groups. The load balancing degree is quantized by the comprehensive load state vector.

[0013] Furthermore, the target density distribution within the overlapping region is calculated as shown in the following formula: , in, , in, Let x be the target density at time t. Let x be the initial target density at position x. This represents the total number of grid cells within the overlapping region. For Gaussian kernel function, It is a spatial position vector. Let i be the center position vector of the i-th grid cell. For adaptive bandwidth parameters, Let be the target's average velocity vector. Based on bandwidth, For reference speed.

[0014] Furthermore, the switching loss is obtained by calculating the mechanical loss between gimbal groups, the field-of-view switching loss, and the tracking accuracy loss, and then performing a weighted combination.

[0015] Compared with existing technologies, the self-diagnosis and predictive maintenance method for the operational status of security system equipment provided by this invention collects the coordinated motion data of each pan-tilt motor in the camera array, calculates the synchronization index of the motion of adjacent camera pan-tilts, and constructs a pan-tilt group control coordination matrix, thereby comprehensively analyzing the coordinated motion characteristics of the pan-tilt devices. Based on this, the load balance of each camera pan-tilt and the optimization of overlapping areas are calculated, and a dynamic task allocation strategy is used to group and control the pan-tilt group, effectively realizing the optimized scheduling of multi-target tracking tasks. This method can not only significantly improve the collaborative working efficiency of the pan-tilt group, but also reduce blind spots and invalid overlapping areas in the monitoring area, optimizing the overall monitoring effect. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a system block diagram of a security system equipment self-diagnosis and predictive maintenance method according to an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the dynamic adjustment and growth of boundary curves based on the characteristics of divergence and curl fields in the self-diagnosis and predictive maintenance method for the operating status of security system equipment according to an embodiment of the present invention. Detailed Implementation

[0018] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0019] Figure 1 This is a system block diagram of a security system equipment self-diagnosis and predictive maintenance method according to an embodiment of the present invention. Figure 1 As shown, the self-diagnosis and predictive maintenance method for the operating status of security system equipment includes: S1: Collect the coordinated motion data of each gimbal motor in the camera array. The coordinated motion data includes the motion timing data between adjacent cameras, gimbal rotation speed data, and load current data. By calculating the synchronization index of the gimbal motion of adjacent cameras, a gimbal group control coordination matrix is ​​constructed. Using the data interface in the driver controller built into the camera pan-tilt unit, the motion timing data in the pan-tilt motion control command is directly read. This motion timing data includes the pan-tilt unit's target position, current position, and motion timestamp. The rotational speed data of the pan-tilt motor is read through the encoder feedback interface built into the pan-tilt motor driver. This rotational speed data includes real-time rotational speed values ​​in both the horizontal and vertical directions. The load current data of the motor is read using the current detection circuit in the pan-tilt driver controller. This load current data records the input current value of the pan-tilt motor during its movement. All the motion timing data, rotational speed data, and load current data are processed through the pan-tilt controller... The standard communication protocol is transmitted to the main control unit of the camera system, which buffers the data according to a sampling period of 100ms. For adjacent cameras connected by the same network switch, their main control units interact with each other via TCP / IP protocol and synchronize time based on Network Time Protocol (NTP). The synchronized cooperative motion data is used to calculate the synchronization index of the pan-tilt-zoom (PTZ) motion of adjacent cameras. The synchronization index is obtained by weighted fusion of position matching degree, speed coordination degree, and load correlation degree. Finally, a PTZ group control cooperative matrix is ​​constructed, which reflects the cooperative motion characteristics between adjacent PTZs.

[0020] The process of using synchronized cooperative motion data to calculate the synchronization index of adjacent camera gimbal movements is as follows: First, the position matching degree is calculated for the motion time series data of adjacent cameras. When two gimbals are moving within the same time window, the position deviation value is calculated using the Euclidean distance between the target position and the current position of the gimbal, and the ratio of this deviation value to a preset reference distance is used as the position matching degree. Second, the speed coordination degree of adjacent gimbals is calculated. After the sampling time of the gimbal rotation speed data is aligned, the speed vectors in the horizontal and vertical directions are calculated respectively. The speed coordination degree is obtained by the cosine value of the angle between the two speed vectors. If the speed vectors are in the same direction, the coordination degree is considered high. The scheduling is the highest priority; next, the correlation of the load current of adjacent PTZs is analyzed. When two PTZs move simultaneously, their load current data are normalized within the same time window, and the load correlation is obtained by calculating the Pearson correlation coefficient of the normalized current waveform; finally, the position matching degree, speed coordination degree and load correlation are weighted and summed to obtain the synchronization index, where the weight coefficients are dynamically adjusted according to the actual motion scenario. When the PTZ performs a tracking task, the weight of speed coordination degree is relatively high, and when the PTZ performs fixed-point monitoring, the weight of position matching degree is relatively high. The obtained synchronization index is used to characterize the degree of cooperative movement between adjacent PTZs.

[0021] The process of constructing a PTZ group control collaboration matrix based on the calculated synchronization index is as follows: First, determine the spatial distribution relationship of the PTZs in the camera array. When the monitoring ranges of two PTZs overlap, they are defined as adjacent PTZ pairs, and a unique identifier is assigned to each adjacent PTZ pair. Second, construct an n×n matrix structure, where n represents the total number of PTZs in the camera array. When the PTZ pair corresponding to the i-th row and j-th column of the matrix is ​​adjacent, the synchronization index of the PTZ pair is filled into the corresponding position in the matrix. If the PTZ pairs are not adjacent, the preset minimum synchronization index value is filled in. Third, since the synchronization index is symmetrical, that is, the synchronization index of PTZ i and PTZ j is equal to the synchronization index of PTZ j and PTZ i, the upper triangular elements of the matrix are mapped to the lower triangular positions. Finally, the diagonal elements represent the state characteristics of the PTZ itself. When the PTZ is in normal working condition, the diagonal elements are filled with the preset maximum synchronization index value. If the PTZ malfunctions or is under maintenance, the preset minimum synchronization index value is filled in.

[0022] S2: Based on the gimbal group control collaboration matrix, calculate the load balance of each camera gimbal during multi-target tracking, wherein the load balance is obtained through the dynamic coupling relationship between gimbal rotation speed and load current, and establish an optimized overlapping area of ​​gimbal motion trajectory. First, the row vector corresponding to each gimbal in the gimbal group control coordination matrix is ​​extracted and converted into polar coordinates, where the amplitude represents the strength of the coordination relationship and the angle represents the spatial positional relationship between the gimbals. Second, the coordination load index of each gimbal is calculated. The coordination load index is obtained by vector superposition of non-zero elements in polar coordinates, and the superposition result forms the coordination load vector. Third, the motion trajectory characteristics of each gimbal in the most recent detection cycle are analyzed. By calculating the curvature change and velocity distribution of the trajectory curve, the dynamic load vector of the gimbal is constructed. Then, the coordination load vector is compared with the dynamic load vector. The load state vectors are synthesized in the same coordinate system to obtain the comprehensive load state vector of the gimbal. The direction of the comprehensive load state vector indicates the trend of load transfer, and the amplitude represents the load level. Finally, based on the comprehensive load state vectors of each gimbal, an optimized overlapping region for the gimbal motion trajectory is established. Specifically, this includes: identifying load accumulation regions and load sparse regions by calculating the divergence field of the comprehensive load state vectors of adjacent gimbals; determining the load circulation direction in the gimbal group based on the curl distribution of the load state vector field; and constructing a dynamic optimization boundary for the gimbal motion trajectory by combining the distribution characteristics of the divergence field and the curl field.

[0023] The process of synthesizing the cooperative load vector and the dynamic load vector in the same coordinate system is as follows: First, a unified load representation space is constructed in a spherical coordinate system, where the radial component represents the load intensity, the azimuth angle represents the horizontal movement direction of the gimbal, and the pitch angle represents the vertical movement direction of the gimbal. Second, a mapping transformation is performed on the cooperative load vector. When the cooperative intensity between adjacent gimbals is strong, the radial component of the cooperative load vector increases accordingly. The azimuth angle is determined by the relative position between the gimbals, and the pitch angle is determined according to the vertical movement correlation of the gimbals. Third, a mapping transformation is performed on the dynamic load vector. When the curvature of the gimbal's trajectory changes drastically, the dynamic load vector... The radial component increases accordingly, the azimuth angle is determined by the current horizontal movement trend of the gimbal, and the pitch angle is determined by the vertical movement trend of the gimbal. Then, using the vector synthesis rule in the spherical coordinate system, the cooperative load vector and the dynamic load vector of the same gimbal are weighted and synthesized. The weight coefficient is determined by calculating the complementary characteristics of the cooperative load and the dynamic load. When the gimbal mainly performs cooperative tracking tasks, the weight of the cooperative load vector is relatively large, and when the gimbal mainly performs independent cruise tasks, the weight of the dynamic load vector is relatively large. Finally, the synthesized comprehensive load state vector is projected onto the Cartesian coordinate system to obtain the load distribution characteristics of the gimbal in three-dimensional space.

[0024] The process of calculating the divergence field of the integrated load state vector of adjacent PTZs is as follows: First, a three-dimensional mesh is established in the monitoring space of the camera array, with each mesh node as a calculation unit. When a mesh node is within the monitoring range of a PTZ, the eight adjacent nodes around that node are constructed as a local calculation domain. Second, for each calculation unit, the integrated load state vector of the PTZs covering that area is extracted. When multiple PTZs simultaneously cover the same calculation unit, the spatial weighting coefficient is calculated based on the distance from the PTZ to the calculation unit. Third, within each local calculation domain, the central difference method is used to calculate the integrated load state vector in three directions. The partial derivatives of the load distribution are corrected using forward or backward differencing methods when the computing unit is at the boundary of the monitoring area. Then, the partial derivatives in the three directions are weighted and superimposed to obtain the divergence value at the computing unit. When the divergence is positive, it indicates that the area is a load dispersion area, and when the divergence is negative, it indicates that the area is a load convergence area. Next, spatial interpolation is performed on the divergence values ​​of all computing units, and a continuous divergence distribution function is constructed using radial basis functions. Finally, load accumulation areas and load sparse areas are identified based on the divergence distribution function. When the spatial gradient of the divergence changes significantly, the area is marked as a key area for load transfer.

[0025] Furthermore, morphological analysis is performed on the divergence distribution function. A sliding block is used to perform local statistical analysis of the divergence values. When the mean divergence value within a local region is negative and the variance is small, the region is initially marked as a load accumulation region. When the mean divergence value within a local region is positive and the variance is small, the region is initially marked as a load sparse region. Next, the spatial gradient field of the divergence distribution function is calculated. By tracing the streamlines of the gradient field, the main path of load transmission is determined. When streamlines converge, local extrema are found along the streamline direction. Third, a feature region growth algorithm is constructed centered on the local extrema. When the divergence value changes of adjacent computational units satisfy the continuity condition, they are included in the current feature region, and the boundary of the feature region is determined iteratively. Finally, temporal correlation analysis is performed on the identified feature regions. When a feature region remains stable within multiple consecutive time windows, it is confirmed as a stable load accumulation region or a load sparse region.

[0026] On the other hand, the load accumulation region and load sparse region of the divergence field are mapped to the monitoring space, and the load circulation direction reflected by the curl field is projected onto the same space. When the characteristic region of the divergence field overlaps with the circulation region of the curl field, the initial boundary point set is determined according to the spatial range of the overlapping region. Secondly, based on the initial boundary point set, a boundary growth algorithm is constructed. When the ratio of divergence value to curl value at the boundary point remains within the dynamic equilibrium range, the boundary is extended along the tangential direction of the curl field. When the absolute value of divergence increases significantly, the boundary direction is adjusted along the divergence gradient direction. Thirdly, the generated boundary curve is smoothed. By calculating the tension and curvature on the boundary curve, the optimal division of the monitoring area is achieved while maintaining the continuity of the boundary.

[0027] The boundary growth algorithm can be expressed by the following formula: , in, , It also includes curve smoothness constraints, as shown in the following formula: , in, Let be the boundary point position vector at time t (unit: m). The time step (in seconds). The boundary growth rate vector (unit: m / s) The tangential spreading rate coefficient (unit: m / s) Normal adjustment rate coefficient (unit: m / s), Let be the tangential unit vector of the curl field. Let be the divergence field scalar (unit: 1 / s). The curl field vector (unit: 1 / s) The divergence-curl ratio threshold (dimensionless) is used. The divergence threshold (unit: 1 / s) The smoothness coefficient (unit: m² / s) The tension coefficient (unit: m) Curvature (unit: 1 / m), Curve parameters (unit: m), It is the unit normal vector (dimensionless).

[0028] S3: Based on the load balancing degree and optimized overlapping area, a dynamic task allocation strategy is adopted to group and control the PTZ group. The dynamic task allocation strategy is based on the principle of minimum motion cost and determines the target tracking task of each PTZ by calculating the switching loss between PTZ groups. First, the load balancing of all PTZs at the current moment is sorted. PTZs with load balancing (quantized by the comprehensive load state vector) within the balanced range are classified as stable groups. When the load balancing of a PTZ exceeds the upper limit of the balanced range, it is classified as an overloaded group. When the load balancing of a PTZ is below the lower limit of the balanced range, it is classified as an idle group. Second, the spatial distribution characteristics of each group of PTZs within the optimized overlapping area are analyzed. When the monitoring areas of the overloaded group PTZs and the idle group PTZs overlap (load accumulation area), the target density distribution within the overlapping area is calculated, and the priority of task migration is determined based on the spatial gradient of the target density. Third, based on the target density... The system dynamically groups PTZs based on their density, prioritizing the allocation of areas with higher target density within the monitoring area to PTZs with better performance. Simultaneously, it considers the spatial distribution of PTZs, ensuring moderate overlap between the monitoring areas of adjacent PTZs to guarantee tracking continuity. Next, it calculates the switching loss between adjacent PTZ groups, including changes in PTZ rotation angle, field-of-view overlap, and target tracking accuracy. When the switching loss is low, task transfer is prioritized. Finally, based on the comprehensive evaluation results of target density distribution and switching loss, the specific tracking task for each PTZ is determined. By adjusting the PTZ grouping scheme and task allocation strategy in real time, the overall tracking performance and load balance of the system are ensured.

[0029] The target density distribution within the overlapping region is calculated using the following formula: , in, , in, Let x be the target density at position x at time t (unit: objects / m²). Let x be the initial target density (unit: targets / m²). This represents the total number of grid cells within the overlapping region. For Gaussian kernel function, This is a spatial location vector (unit: m). Let be the center position vector of the i-th grid cell (unit: m). For adaptive bandwidth parameters (unit: meters), Let be the target's average velocity vector (unit: m / s). The base bandwidth (unit: meters). Reference speed (unit: m / s).

[0030] The specific process for calculating the switching loss between adjacent gimbal groups is as follows: First, the spatial positional relationship between adjacent gimbal groups is analyzed, and the distance matrix and angular difference between the two groups are calculated. When the relative positional relationship between the gimbal groups changes, the position parameters are updated in real time. Second, the mechanical loss during gimbal rotation is calculated, including the change in angular velocity, acceleration, and energy consumption required for the gimbal to move from its current position to the target position. When the gimbal needs to rotate rapidly at a large angle, the mechanical loss value is larger. Third, the changing characteristics of the overlapping field of view are analyzed, and the image quality change during the process of the target moving from one gimbal's field of view to another is calculated. When the field-of-view overlap is small or the imaging resolution difference is large, the field-of-view switching loss increases. Then, the change in target tracking accuracy is evaluated, including the target's position deviation on the image plane, velocity estimation error, and the probability of transient target loss. When the tracking algorithm needs to be reinitialized or adapted to new imaging conditions, the tracking accuracy loss increases. Finally, the mechanical loss, field-of-view switching loss, and tracking accuracy loss are weighted and combined to obtain a comprehensive switching loss value. The weighting coefficients are dynamically adjusted according to the needs of the actual application scenario. This switching loss value serves as an important basis for task allocation decisions, reducing the overall switching overhead of the system by optimizing the task transfer timing. The comprehensive switching loss value can be expressed by the following formula: , in, The total switching loss value (unit: J) Angular velocity loss coefficient (unit: s), The acceleration loss coefficient (unit: kg·m) The moment of inertia of the gimbal (unit: kg·m²) The change in angular velocity (unit: rad / s) The change in acceleration (unit: m / s²) The field of view overlap (dimensionless) The target position deviation (unit: meters). The imaging quality factor (dimensionless) Target tracking confidence (dimensionless).

[0031] For example, multiple pan-tilt cameras are deployed in a stadium. When a match is about to end and a large number of spectators are leaving, a pan-tilt camera near the main exit area detects a sharp increase in target density within its monitoring area, while the target density in the areas of the two adjacent pan-tilt cameras is relatively low. The system first calculates the target density distribution in the overlapping areas between the high-load pan-tilt camera and its neighbors. It finds that the target density in one overlapping area shows a clear flow trend, while the target density in the other overlapping area is relatively stable. Next, the system calculates the switching loss: because one of the adjacent pan-tilt cameras is tracking a key target at a relatively high speed, it needs to rotate at a large angle to help share the load. Furthermore, due to the low overlap with the high-load pan-tilt camera's field of view, this results in a large switching loss, leading to a high overall switching loss value. Conversely, the other adjacent pan-tilt camera is relatively stationary, requiring only a small rotation, and has a high overlap with the high-load pan-tilt camera's field of view, resulting in a lower overall switching loss value. Based on the above calculation results, the system ultimately decided to use a gimbal with lower switching loss to assist in task sharing. The system transferred some targets in the coverage area of ​​the high-load gimbal to the side of the assisting gimbal for tracking, thereby achieving load balancing with low switching loss.

[0032] S4: Adaptively smooth the motion trajectory of each gimbal group, and generate a gimbal wear assessment report by analyzing the cooperative motion characteristics of the gimbal group. The cooperative motion characteristics include the motion consistency of the gimbals within the group and the task switching frequency between groups.

[0033] The system collects motion parameters such as angular velocity, angular acceleration, and motor torque of each gimbal group in real time during operation. When gimbal movement exhibits jitter or sudden speed changes, the motion trajectory is immediately smoothed and corrected. Secondly, it statistically analyzes the start-stop frequency, motion duration, and load of the gimbal groups during collaborative tracking. If the usage intensity of a particular gimbal is too high, its task allocation is adjusted to reduce the load. Thirdly, it monitors the status of the mechanical transmission chain of the gimbal groups, including observing noise during gear meshing, the smoothness of bearing operation, and motor heating, recording abnormal conditions and addressing them promptly. Next, it records environmental factors such as temperature, humidity, and dust, analyzing the impact of these factors on gimbal wear. Simultaneously, it monitors the vibration characteristics and noise levels of the gimbal groups during operation, promptly repairing any abnormalities. Finally, based on the cumulative operating time, load records, fault history, and maintenance status of the gimbal groups, it statistically analyzes various data to generate an assessment report containing a description of wear status, component wear levels, potential fault warnings, and maintenance recommendations, providing a basis for the daily maintenance and management of the gimbal groups.

[0034] In summary, the self-diagnosis and predictive maintenance method for the operational status of security system equipment based on the embodiments of the present invention has been clarified. By collecting the coordinated motion data of each pan-tilt motor in the camera array, calculating the synchronization index of the pan-tilt motion of adjacent cameras, and constructing a pan-tilt group control coordination matrix, the coordinated motion characteristics of the pan-tilt devices can be comprehensively analyzed. Based on this, the load balance of each camera pan-tilt and the optimized overlapping area are calculated, and a dynamic task allocation strategy is used to group and control the pan-tilt group, effectively achieving optimized scheduling of multi-target tracking tasks. This method can not only significantly improve the collaborative working efficiency of the pan-tilt group but also reduce blind spots and invalid overlapping areas in the monitoring area, optimizing the overall monitoring effect.

[0035] Here, those skilled in the art will understand that the specific operations of each step in the above-described method for self-diagnosis and predictive maintenance of security system equipment operating status have been referenced above. Figure 1 and Figure 2 The description of the self-diagnosis and predictive maintenance methods for the operating status of security system equipment is detailed here, and therefore, its repeated description will be omitted.

[0036] In summary, the self-diagnosis and predictive maintenance method for the operational status of security system equipment based on the embodiments of the present invention has been clarified. By collecting the coordinated motion data of each pan-tilt motor in the camera array, calculating the synchronization index of the pan-tilt motion of adjacent cameras, and constructing a pan-tilt group control coordination matrix, the coordinated motion characteristics of the pan-tilt devices can be comprehensively analyzed. Based on this, the load balance of each camera pan-tilt and the optimized overlapping area are calculated, and a dynamic task allocation strategy is used to group and control the pan-tilt group, effectively achieving optimized scheduling of multi-target tracking tasks. This method can not only significantly improve the collaborative working efficiency of the pan-tilt group but also reduce blind spots and invalid overlapping areas in the monitoring area, optimizing the overall monitoring effect.

Claims

1. A security system device operation state self-diagnosis and predictive maintenance method, characterized by, The method comprises the following steps: Collecting the cooperative motion data of each PTZ camera motor in the camera array, constructing a PTZ group control cooperation matrix by calculating the synchronization index of the motion of adjacent PTZ cameras, and calculating the load balancing degree of each PTZ camera in the multi-target tracking process based on the PTZ group control cooperation matrix, and establishing an optimized overlapping area of the PTZ motion trajectory; According to the load balancing degree and the optimized overlapping area, a dynamic task allocation strategy is adopted to control the PTZ group, and the target tracking task of each PTZ is determined by calculating the switching loss between the PTZ groups; The motion trajectory of each PTZ group is adaptively smoothed, and a PTZ wear evaluation report is generated by analyzing the cooperative motion characteristics of the PTZ group. The calculation of the synchronization index comprises the following steps:

2. The security system device operational status self-diagnostic and predictive maintenance method of claim 1, wherein, The position matching degree is calculated based on the ratio of the PTZ position deviation to the reference distance; The speed vector in the horizontal and vertical directions is calculated based on the PTZ rotation speed data, and the speed coordination degree is obtained by the cosine value of the included angle between the two speed vectors; The load correlation degree is obtained by calculating the Pearson correlation coefficient of the normalized current waveforms of the two PTZs; The position matching degree, the speed coordination degree and the load correlation degree are summed by weighting to obtain the synchronization index. The calculation of the load balancing degree is to synthesize the cooperative load vector and the dynamic load vector of the PTZ in the spherical coordinate system to obtain the comprehensive load state vector.

3. The security system device operational status self-diagnostic and predictive maintenance method of claim 2, wherein, The cooperative load vector is obtained by vector superposition of non-zero elements in polar coordinates. The dynamic load vector is obtained by calculating the trajectory curvature change and speed distribution characteristics of the PTZ in the latest detection period. The establishment of the optimized overlapping area of the PTZ motion trajectory is to identify the load distribution characteristics by calculating the divergence field of the comprehensive load state vectors of adjacent PTZs, and to determine the dynamic optimization boundary of the PTZ motion trajectory by combining the circulation direction of the vorticity field under the condition of satisfying the boundary continuity.

4. The security system device operational status self-diagnostic and predictive maintenance method of claim 3, wherein, The identification of the load distribution characteristics is to locally count the divergence value by using a sliding block, and to mark the region with negative divergence mean value and small variance as a load accumulation region, and to mark the region with positive divergence mean value and small variance as a load sparse region.

5. The security system device operational status self-diagnostic and predictive maintenance method of claim 4, wherein, The boundary continuity is determined based on the boundary growth algorithm; 6. The security system device operational status self-diagnostic and predictive maintenance method of claim 4, wherein, The boundary growth algorithm expands the boundary along the tangent of the vorticity field within the range of dynamic balance according to the ratio of the divergence value to the vorticity value at the boundary point, and adjusts the boundary direction along the divergence gradient direction when the absolute value of the divergence significantly increases. The boundary growth algorithm is as follows:

7. The security system device operational status self-diagnostic and predictive maintenance method of claim 6, wherein, Wherein, ， The dynamic task allocation strategy is to group the PTZs according to the load balancing degree, and then to optimize the task allocation scheme based on the target density distribution of the overlapping area and the switching loss between the PTZ groups; ， wherein, is a position vector of a boundary point at time t, is a time step, is a velocity vector of the boundary growth, is a tangential spreading rate coefficient, is a normal adjustment rate coefficient, is a tangential unit vector of the curl field, is a scalar of the divergence field, is a vector of the curl field, is a divergence to curl ratio threshold value, is a divergence threshold value.

8. The security system device operational status self-diagnostic and predictive maintenance method of claim 1, wherein, The load balancing degree is quantified by the comprehensive load state vector. The target density distribution in the overlapping area is calculated as follows:

9. The security system device operational status self-diagnostic and predictive maintenance method of claim 8, wherein, Wherein, ， The switching loss is obtained by calculating and weighting the mechanical loss, the field of view switching loss and the tracking accuracy loss between the PTZ groups. ， wherein, is the target density at position x at time t, is the initial target density at position x, is the total number of grid cells in the overlap region, is the Gaussian kernel function, is the spatial position vector, is the center position vector of the i-th grid cell, is the adaptive bandwidth parameter, is the average velocity vector of the target, is the base bandwidth, is the reference velocity.

10. The security system device operational status self-diagnostic and predictive maintenance method of claim 8, wherein, ​

Images