Fault self-checking recovery system for video monitoring system in sports place

Abstract

The invention belongs to the field of monitoring fault self-checking, relates to a data analysis technology, and particularly relates to a fault self-checking recovery system for a video monitoring system in a sports place, which is used for solving the problem that in the prior art, the angle of a camera generating deviation cannot be corrected in time according to a self-checking result. Comprising an offset detection module, an offset calibration module and a habit analysis module which are connected in sequence, and the offset detection module, the offset calibration module and the habit analysis module are all in communication connection with a database; according to the method, the problem of target loss caused by camera offset of a sports place monitoring system is effectively solved, and automatic detection and rapid correction of visual angle offset are realized by establishing a comparison mechanism of the reference coordinates and the real-time coordinates. The technical means is particularly suitable for sports places with frequent mechanical vibration, such as a football field and a basketball court, potential monitoring blind areas can be eliminated before the equipment is not completely invalid, and the reliability and the maintenance efficiency of a video monitoring system are remarkably improved.

Description

Technical Field

[0001] The invention belongs to the field of monitoring fault self-detection and relates to data analysis technology, and in particular to a fault self-detection and recovery system for a video monitoring system in a sports venue. Background Art

[0002] The sports venue monitoring system is a comprehensive management system based on intelligent perception, video analysis and Internet of Things technologies. It is mainly used to monitor the safety, equipment status, personnel activities and environmental parameters of sports venues in real time to improve management efficiency, ensure personnel safety and optimize operations.

[0003] The invention patent with publication number CN107272637B discloses a video surveillance system fault self-detection and self-recovery control system and method. The self-detection and recovery control system can realize the display, recording, prompting, etc. of the operation and fault status of each device in the entire system, so as to achieve the functions of monitoring the safe and stable operation status of the system and judging the fault point, thereby reducing the workload and difficulty of system equipment maintenance; however, when the system is used in sports venues with high failure rates and offset rates, it is impossible to correct the angle of the camera that has offset in time according to the self-detection results, nor is it possible to monitor the habitual offset of the camera in a specific environment through the offset self-detection results and the angle correction process, resulting in low video surveillance viewing angle accuracy and maintenance efficiency. Summary of the Invention

[0004] The purpose of the present invention is to provide a fault self-diagnosis and recovery system for a sports venue video surveillance system, which is used to solve the problem that the existing technology cannot timely correct the angle of the camera that has deviated according to the self-diagnosis results; The technical problem to be solved by the present invention is: how to provide a fault self-checking and recovery system for a video surveillance system in a sports venue, which can promptly correct the angle of a camera that has deviated according to the self-checking result.

[0005] The purpose of the present invention can be achieved through the following technical solutions: A fault self-detection and recovery system for a sports venue video surveillance system includes an offset detection module, an offset calibration module, and a habit analysis module connected in sequence, wherein the offset detection module, the offset calibration module, and the habit analysis module are all communicatively connected to a database; The offset detection module is used to perform offset detection and analysis on a static camera: the static camera is marked as a detection object, the angle of the detection object is adjusted according to the shooting requirements, and after the angle adjustment is completed, an image is captured by the detection object to obtain a standard image, and the standard image is used to determine whether the detection object has lost offset or offset abnormality; The offset calibration module is used to perform offset calibration analysis on a static camera: generate an offset sequence according to the result of the offset detection analysis, and perform angle adjustment calibration on the detection object through the offset sequence; The habit analysis module is used to analyze the deviation habits of the sports venue video surveillance system.

[0006] Furthermore, it is determined whether the detection object has lost offset: several static objects in the standard image are extracted and marked as labeled object i, i=1, 2,…,n, where n is a positive integer, the coordinates of the standard object i in the standard image are obtained and marked as the reference coordinates JZi of the standard object i, JZi=(JXi, JYi), the video shot when the detection object is running is decomposed into analysis images frame by frame, the labeled object i in the analysis image is extracted and it is determined whether all labeled objects i are extracted: if so, it is determined that the detection object has no lost offset; if not, it is determined that the detection object has lost offset, and basic correction processing is performed on the detection object.

[0007] Furthermore, the specific process of performing basic correction processing on the detection object includes: continuously and randomly adjusting the shooting angle of the detection object until all the labeled objects i are extracted from the captured image, and marking the current captured image as the analysis image; if the basic correction cannot be completed, a defacement processing signal is generated and sent to the administrator's mobile phone terminal.

[0008] Furthermore, the benchmark offset analysis is performed on the analysis image: the coordinates of the extracted annotation object i are marked as real-time coordinates SSi, SSi=(SXi,SYi), and the formula The deviation coefficient PY of the analysis image is obtained, and whether the detection object has deviation abnormality is determined by the deviation coefficient PY.

[0009] Furthermore, the specific process of determining whether the detection object has an offset abnormality includes: retrieving the offset threshold PYmax through the database, and comparing the offset coefficient PY with the offset threshold: if the offset coefficient PY is less than the offset threshold PYmax, it is determined that the detection object does not have an offset abnormality; if the offset coefficient PY is greater than or equal to the offset threshold PYmax, it is determined that the detection object has an offset abnormality, generating an offset calibration signal and sending the offset calibration signal to the offset calibration module.

[0010] Furthermore, the generation process of the offset sequence includes: Obtain the offset PLi of the labeled object i, and arrange the labeled objects i in descending order according to the value of the offset PLi to obtain an offset sequence.

[0011] Furthermore, the specific process of angular adjustment calibration of the detection object includes: extracting the lateral offset feature HTi and the longitudinal offset feature ZTi of the marked object i ranked first in the offset sequence, wherein the lateral offset feature HTi is the difference between the horizontal coordinate value of the reference coordinate JZi of the marked object i and the horizontal coordinate value of the real-time coordinate SSi, and the longitudinal offset feature ZYi is the difference between the vertical coordinate value of the reference coordinate JZi of the marked object i and the vertical coordinate value of the real-time coordinate SSi; converting the lateral offset feature HTi and the longitudinal offset feature ZTi into the calibration angle of the detection object, and adjusting the detection object according to the calibration angle.

[0012] Furthermore, the process of calibrating the angle adjustment of the detection object also includes: after the adjustment is completed, marking the captured image of the detection object as an adjustment image, obtaining the offset coefficient PY of the adjustment image and determining whether there is an offset abnormality in the adjustment image: if not, the calibration is completed; if so, regenerating the offset sequence and calibration angle according to the coordinates of the marked object i in the adjustment image, and performing a second adjustment on the detection object through the calibration angle, and so on, until there is no offset abnormality in the detection object.

[0013] Furthermore, the specific process of the habit analysis module analyzing the offset habits of the sports venue video surveillance system includes: generating an analysis cycle, marking the lateral offset feature HTi and the longitudinal offset feature ZTi corresponding to the first calibration angle generated when the detection object performs offset calibration analysis within the analysis cycle as lateral habit value HX and longitudinal habit value ZX respectively, performing variance calculation on all lateral habit values HX within the analysis cycle to obtain a lateral habit coefficient, and performing variance calculation on all longitudinal habit values ZX within the analysis cycle to obtain a longitudinal habit coefficient; and judging whether the detection object has lateral offset habits and longitudinal offset habits through the lateral habit coefficient and the longitudinal habit coefficient.

[0014] Furthermore, the specific process of determining whether the detected object has the habit of lateral displacement and longitudinal displacement includes: obtaining the habit threshold through the database; comparing the lateral habit coefficient and the longitudinal habit coefficient with the habit threshold respectively; if the lateral habit coefficient is less than the habit threshold, then it is determined that the detected object has the habit of lateral displacement, and a lateral displacement processing signal is generated and sent to the mobile phone terminal of the administrator; if the longitudinal habit coefficient is less than the habit threshold, then it is determined that the detected object has the habit of longitudinal displacement, and a longitudinal displacement processing signal is generated and sent to the mobile phone terminal of the administrator; otherwise, it is determined that the detected object does not have the habit of displacement.

[0015] The present invention has the following beneficial effects: This invention effectively solves the problem of target loss caused by camera offset in sports venue monitoring systems. By establishing a comparison mechanism between reference coordinates and real-time coordinates, it achieves automated detection and rapid correction of viewing angle offset. This technology is particularly suitable for sports venues with frequent mechanical vibrations, such as football fields and basketball halls. It can eliminate potential monitoring blind spots before equipment fails completely, significantly improving the reliability and maintenance efficiency of video surveillance systems. The present invention can address the complex situation of multi-target offset in sports venue monitoring scenarios by establishing a calibration priority mechanism based on quantitative sorting, ensuring that the offset of key areas is corrected first during camera angle adjustment, improving the accuracy and execution efficiency of calibration operations, and thus quickly restoring the normal coverage range of the monitoring angle. The present invention can identify the habitual offset patterns formed by cameras in specific usage scenarios, help managers locate the causes of repetitive failures, reduce frequent calibration operations caused by regular offsets, and optimize maintenance strategies to extend the service life of equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 This is a system block diagram of Embodiment 1 of the present invention; Figure 2 This is a flow chart of the method of embodiment 2 of the present invention. DETAILED DESCRIPTION

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0019] In existing technologies, video surveillance systems in sports venues have long faced the problem of perspective deviation caused by camera offset. Due to the frequent movement of equipment, human activity, and environmental vibrations in sports venues, fixed cameras are susceptible to external forces and cause angular deviation. While traditional monitoring systems can detect the operating status of equipment, they lack a dynamic detection mechanism for camera perspective deviation, making it difficult to trigger calibration procedures in a timely manner. When cameras habitually offset due to long-term mechanical wear or a loose mounting base, existing systems are unable to identify these regular anomalies, resulting in maintenance personnel having to perform repeated manual corrections, significantly increasing operation and maintenance costs.

[0020] To address these issues, the R&D team observed that camera offsets in sports venues exhibit both sudden and cumulative characteristics. Sudden offset is typically caused by external impact and requires rapid detection and correction. Cumulative offset, however, stems from equipment aging and manifests as periodic angular deviations. Traditional, single-dimensional fault detection methods are unable to distinguish between these two types of offset characteristics. By breaking down the detection process into three phases: real-time anomaly detection, dynamic calibration execution, and long-term behavioral analysis, a hierarchical processing mechanism was established. First, a standard image baseline was established to identify sudden offsets by comparing them to real-time images. Next, a prioritized calibration sequence was generated based on the extent of the offset. Finally, long-term data analysis was used to capture periodic offset patterns, forming a basis for preventative maintenance.

[0021] Example 1: Figure 1 As shown, a fault self-detection and recovery system for a sports venue video surveillance system includes an offset detection module, an offset calibration module, and a habit analysis module, all connected in sequence. The offset detection module, offset calibration module, and habit analysis module are all communicatively connected to a database. The offset detection module is a detection unit that performs a baseline comparison using a preset standard image. Specifically, this can be achieved using an image feature point matching algorithm, such as an ORB feature point detector, to extract the coordinates of key points in the image and establish a coordinate mapping relationship between the standard image and the real-time image. This component solves the problem of instantaneous identification of sudden offsets by establishing a visual baseline. The offset calibration module is a control unit that performs corrective actions based on offset priority. Specifically, it can use a PID control algorithm to adjust the pan / tilt rotation angle. By prioritizing the object with the largest offset, the view angle of key monitoring areas is restored first. The habit analysis module is an analysis unit that performs trend modeling on historical calibration data. Specifically, it can use time series analysis methods, such as the ARIMA model to identify periodic offset patterns. This module solves the problem of predicting cumulative offsets through data mining. The database is a storage unit that stores standard image data and calibration records. Specifically, it can use a distributed time series database to achieve high-frequency data access, providing a unified data support platform for the three functional modules.

[0022] The offset detection module is used to perform offset detection and analysis on static cameras: the static camera is marked as the detection object. It should be noted that the static camera is a camera with a fixed shooting angle requirement, and does not include a camera that automatically adjusts the shooting angle for target tracking; the angle of the detection object is adjusted according to the shooting requirements. After the angle adjustment is completed, the image is captured by the detection object to obtain a standard image, and several static objects in the standard image are extracted and marked as labeled objects i, i=1, 2,…,n, n is a positive integer, the coordinates of the standard object i in the standard image are obtained and marked as the reference coordinates JZi of the standard object i, JZi=(JXi, JYi), the video shot when the detection object is running is decomposed into analysis images frame by frame, the labeled objects i in the analysis image are extracted and it is determined whether all labeled objects i have been extracted: if so, the detection object is determined to be The detected object does not have a missing offset; if not, it is determined that the detected object has a missing offset, and basic correction processing is performed on the detected object: the shooting angle of the detected object is continuously and randomly adjusted until all the labeled objects i are extracted from the captured image, and the current captured image is marked as the analysis image; if the basic correction cannot be completed, a contamination processing signal is generated and the contamination processing signal is sent to the administrator's mobile terminal; when the labeled object i is detected to be missing, the system starts the angle adjustment mechanism, such as randomly changing the horizontal and pitch angles at a frequency of three times per second, and performing image analysis immediately after each adjustment; if all the labeled objects are successfully captured within the preset number of times (for example, 30 times), the current picture is used as the benchmark for continued monitoring; if the adjustment threshold is exceeded and the correction is still not completed, it is determined that a physical fault exists, and an alarm message containing the device number and fault type is automatically sent to the preset terminal.

[0023] Among them, the labeled object i refers to a static object extracted from the standard image, which can be achieved by using an image recognition algorithm to mark objects in fixed positions, and is used to provide a benchmark reference for subsequent coordinate comparison. The reference coordinate JZi refers to the position data of the labeled object i in the standard image, which can be achieved by converting the pixel position into spatial coordinates through a coordinate system conversion algorithm, and is used to establish a spatial positioning benchmark when the detection object is operating normally. Analyzing the image refers to decomposing the video stream into single-frame images in chronological order, which can be achieved by using video decoding technology, and is used to detect the integrity and position offset of the labeled object frame by frame. Basic correction processing refers to random angle adjustment of the detection object, which can be achieved by controlling the pan-tilt rotation mechanism, and is used to quickly restore the capture capability of the target object when a loss offset occurs.

[0024] Specifically, when the detection object is running, the video stream it captures is disassembled into a sequence of single-frame images. For each frame of the image, the system attempts to extract all pre-marked annotated objects i. If all annotated objects i are successfully extracted, it means that the camera has not offset and caused the target to be lost; if some annotated objects i cannot be extracted, it is determined that the camera has a perspective loss problem. At this time, the system triggers the basic correction processing, and randomly adjusts the camera angle until all annotated objects i reappear in the shooting picture. This process uses a passive trigger mechanism and only starts the correction operation when the target is detected to be lost. Perform benchmark offset analysis on the analysis image: mark the coordinates of the extracted annotation object i as real-time coordinates SSi, SSi=(SXi,SYi); through the formula Obtain the offset coefficient PY of the analyzed image, retrieve the offset threshold PYmax from the database, and compare the offset coefficient PY with the offset threshold: if the offset coefficient PY is less than the offset threshold PYmax, it is determined that the detection object does not have an offset abnormality; if the offset coefficient PY is greater than or equal to the offset threshold PYmax, it is determined that the detection object has an offset abnormality, generate an offset calibration signal, and send the offset calibration signal to the offset calibration module; The offset coefficient is a quantitative indicator calculated by comparing the real-time coordinates of the annotated object with the reference coordinates. This can be achieved using a weighted coordinate difference algorithm and objectively reflects the severity of the camera's viewing angle offset. The offset threshold is a pre-stored threshold in the database. This threshold can be set based on historical calibration data or device performance parameters. For example, it can be set to a range of 5%-8% of the pixel coordinate difference as a critical condition for triggering calibration. The offset calibration signal is a control instruction containing the device number and abnormal coordinate data. This signal can be implemented using JSON or a binary encoding protocol and is used to drive the calibration module to perform angle correction operations.

[0025] Specifically, after calculating the offset coefficient, the system automatically accesses a database to retrieve preset offset criteria. If the offset exceeds the allowable range, a calibration instruction generation mechanism is immediately triggered. For example, in a stadium monitoring scenario, when the cumulative offset of the goal area coordinates exceeds a safety threshold, the system initiates an automatic calibration process without manual intervention. A data verification mechanism is used during calibration signal transmission to ensure instruction integrity. Upon receiving the signal, the calibration module immediately executes the corresponding mechanical adjustment.

[0026] The offset calibration module is used to perform offset calibration analysis on static cameras: Obtain the offset PLi of the labeled object i, arrange the labeled objects i in descending order according to the values of the offset PLi to obtain an offset sequence, extract the lateral offset feature HTi and longitudinal offset feature ZTi of the labeled object i ranked first in the offset sequence, where PTi=(JXi-SXi), ZTi=(JYi-SYi); convert the lateral offset feature HTi and the longitudinal offset feature ZTi into the calibration angle of the detection object, adjust the detection object according to the calibration angle, mark the captured image of the detection object as the adjustment image after the adjustment is completed, obtain the offset coefficient PY of the adjustment image and determine whether the adjustment image has an offset anomaly: if not, the calibration is completed; if so, regenerate the offset sequence and calibration angle according to the coordinates of the labeled object i in the adjustment image, perform a second adjustment on the detection object using the calibration angle, and so on, until the detection object has no offset anomaly; The offset PLi is the difference in spatial distance between the coordinates of annotated object i in the real-time image and the reference coordinates, reflecting the degree of offset for each annotated object i. The offset sequence is a list of annotated objects sorted by offset from largest to smallest. This can be achieved using a quick sort algorithm, prioritizing objects with larger offsets during calibration.

[0027] Specifically, when an abnormal camera offset is detected, the system calculates the difference between the real-time coordinates and the reference coordinates of each annotated object i. The system then calculates the square root of the sum of the squares of the lateral and longitudinal offsets and multiplies it by a weight coefficient to obtain the offset value PLi for each annotated object. For example, the weight coefficient can be a value set based on the importance of the annotated object's position in the surveillance image, with annotated objects located in the center of the image being given a higher weight. All annotated objects are then sorted in descending order of their PLi values to generate an offset sequence, allowing the calibration module to prioritize annotated objects with large offsets or critical locations.

[0028] The habit analysis module is used to analyze the offset habits of the video surveillance system of the sports venue: generate an analysis cycle, mark the lateral offset feature HTi and the longitudinal offset feature ZTi corresponding to the first calibration angle generated when the detection object performs offset calibration analysis within the analysis cycle as lateral habit value HX and longitudinal habit value ZX respectively, perform variance calculation on all lateral habit values HX within the analysis cycle to obtain the lateral habit coefficient, and perform variance calculation on all longitudinal habit values ZX within the analysis cycle to obtain the longitudinal habit coefficient; obtain the habit threshold through the database: compare the lateral habit coefficient and the longitudinal habit coefficient with the habit threshold respectively: if the lateral habit coefficient is less than the habit threshold, it is determined that the detection object has a lateral offset habit, generate a lateral offset processing signal and send the lateral offset processing signal to the mobile phone terminal of the manager; if the longitudinal habit coefficient is less than the habit threshold, it is determined that the detection object has a longitudinal offset habit, generate a longitudinal offset processing signal and send the longitudinal offset processing signal to the mobile phone terminal of the manager; otherwise, it is determined that the detection object does not have an offset habit.

[0029] The analysis period is a pre-set fixed time period used to collect data on offset calibration data. This can be implemented on a daily, weekly, or monthly basis, and is used to periodically assess camera offset trends. The lateral habit value refers to the offset of the lateral angle during the initial adjustment of each calibration process. This value can be calculated by calculating the difference between the horizontal coordinates of the reference coordinates and the real-time coordinates, reflecting the camera's initial lateral offset characteristics. The longitudinal habit value refers to the offset of the longitudinal angle during the initial adjustment of each calibration process. This value can be calculated by calculating the difference between the vertical coordinates of the reference coordinates and the real-time coordinates, reflecting the camera's initial longitudinal offset characteristics. The lateral habit coefficient is the calculated variance of the lateral habit value. It quantifies the stability of the lateral offset by measuring the dispersion of all lateral habit values within the statistical period. The longitudinal habit coefficient is the calculated variance of the longitudinal habit value. This value can be measured by measuring the dispersion of all longitudinal habit values within the statistical period. The habit threshold is a critical value used to determine whether the offset is regular. This threshold can be set based on historical data or experimental testing to distinguish between random and systematic offsets.

[0030] Specifically, during the analysis cycle, the initial adjusted lateral and longitudinal offsets are recorded as habitual values each time a calibration operation is performed. After the cycle ends, the variance of the lateral and longitudinal habitual values is calculated. If the variance calculation result is lower than the habitual threshold, it indicates that the offset has shown a stable trend with low discreteness within the cycle, and habitual offset is determined to exist. For example, if the lateral habitual coefficient is continuously lower than the threshold, it means that the camera frequently experiences similar degrees of lateral offset within a fixed time period, which may be caused by a loose mounting structure or environmental vibration. At this time, the corresponding offset processing signal is generated and sent to the management personnel, prompting targeted maintenance.

[0031] Example 2: Figure 2 As shown, a fault self-detection and recovery method for a sports venue video surveillance system includes the following steps: Step 1: Perform offset detection and analysis on the static camera: Mark the static camera as the detection object, generate a standard image based on the shooting requirements, extract several static objects in the standard image and mark them as labeled objects i, decompose the video captured when the detection object is running into analysis images frame by frame, compare the coordinates of the analysis image with the labeled objects i in the standard image, and use the comparison results to determine whether the detection object has offset; Step 2: Perform offset calibration analysis on the static camera: Generate an offset sequence, extract the lateral offset feature HTi and longitudinal offset feature ZTi of the labeled object i ranked first in the offset sequence, and adjust the detection object based on the lateral offset feature HTi and longitudinal offset feature ZTi; Step 3: Analyze the offset habits of the video surveillance system in sports venues: Generate an analysis cycle, and numerically calculate the lateral offset feature HTi and the longitudinal offset feature ZTi corresponding to the first calibration angle generated when the detection object is subjected to offset calibration analysis within the analysis cycle to obtain the lateral offset coefficient and the longitudinal offset coefficient. Compare the lateral habit coefficient and the longitudinal habit coefficient with the habit threshold respectively, and determine whether the detection object has a lateral offset habit or a longitudinal offset habit based on the comparison results.

[0032] A fault self-checking and recovery system for a video surveillance system in a sports venue, during operation, marks a static camera as a detection object, generates a standard image according to shooting requirements, extracts several static objects in the standard image and marks them as labeled objects i, decomposes the video shot when the detection object is running into analysis images frame by frame, compares the coordinates of the analysis image with the labeled object i in the standard image, and determines whether the detection object has offset based on the comparison results; generates an offset sequence, extracts the lateral offset feature HTi and the longitudinal offset feature ZTi of the labeled object i ranked first in the offset sequence, and adjusts the detection object based on the lateral offset feature HTi and the longitudinal offset feature ZTi; generates an analysis cycle, numerically calculates the lateral offset feature HTi and the longitudinal offset feature ZTi corresponding to the first calibration angle generated during offset calibration analysis of the detection object within the analysis cycle, and obtains lateral and longitudinal offset coefficients; compares the lateral habit coefficient and the longitudinal habit coefficient with habit thresholds, respectively, and determines whether the detection object has a lateral or longitudinal offset habit based on the comparison results.

[0033] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.

[0034] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0035] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A fault self-diagnosis and recovery system for a sports venue video surveillance system, characterized in that: It includes an offset detection module, an offset calibration module and a habit analysis module connected in sequence, wherein the offset detection module, the offset calibration module and the habit analysis module are all connected to the database for communication; The offset detection module is used to perform offset detection and analysis on a static camera: the static camera is marked as a detection object, the angle of the detection object is adjusted according to the shooting requirements, and after the angle adjustment is completed, an image is captured by the detection object to obtain a standard image, and the standard image is used to determine whether the detection object has lost offset or offset abnormality; The offset calibration module is used to perform offset calibration analysis on a static camera: generate an offset sequence according to the result of the offset detection analysis, and perform angle adjustment calibration on the detection object through the offset sequence; The habit analysis module is used to analyze the deviation habits of the sports venue video surveillance system.

2. A fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 1, characterized in that: Determine whether the detection object has lost offset: extract several static objects in the standard image and mark them as labeled objects i, i = 1, 2, ..., n, where n is a positive integer. Obtain the coordinates of the standard object i in the standard image and mark them as the reference coordinates JZi of the standard object i. Decompose the video shot when the detection object is running into analysis images frame by frame. Extract the labeled objects i in the analysis images and determine whether all labeled objects i have been extracted. If so, it is determined that the detection object has not lost offset. If not, it is determined that the detection object has a missing offset and basic correction processing is performed on the detection object.

3. A fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 2, characterized in that: The specific process of performing basic correction processing on the detection object includes: continuously and randomly adjusting the shooting angle of the detection object until all the labeled objects i are extracted from the captured image, and marking the current captured image as the analysis image; if the basic correction cannot be completed, a defacement processing signal is generated and sent to the administrator's mobile terminal.

4. A fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 3, characterized in that: Perform baseline offset analysis on the analysis image: mark the coordinates of the extracted labeled object i as the real-time coordinates SSi, and calculate the offset coefficient PY of the analysis image by comparing the baseline coordinates JZi and the real-time coordinates SSi of all labeled objects i. Use the offset coefficient PY to determine whether the detected object has an offset anomaly.

5. The fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 4, characterized in that: The specific process of determining whether the detection object has an offset abnormality includes: retrieving the offset threshold PYmax through the database, and comparing the offset coefficient PY with the offset threshold: if the offset coefficient PY is less than the offset threshold PYmax, it is determined that the detection object does not have an offset abnormality; if the offset coefficient PY is greater than or equal to the offset threshold PYmax, it is determined that the detection object has an offset abnormality, generating an offset calibration signal and sending the offset calibration signal to the offset calibration module.

6. A fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 5, characterized in that: The generation process of the offset sequence includes: obtaining the offset PLi of the annotation object i by numerically calculating the reference coordinate JZi and the real-time coordinate SSi of the annotation object i, and arranging the annotation object i in descending order according to the value of the offset PLi to obtain the offset sequence.

7. A fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 6, characterized in that: The specific process of angular adjustment calibration of the detection object includes: extracting the lateral offset feature HTi and the longitudinal offset feature ZTi of the marked object i ranked first in the offset sequence, wherein the lateral offset feature HTi is the difference between the horizontal coordinate value of the reference coordinate JZi of the marked object i and the horizontal coordinate value of the real-time coordinate SSi, and the longitudinal offset feature ZYi is the difference between the vertical coordinate value of the reference coordinate JZi of the marked object i and the vertical coordinate value of the real-time coordinate SSi; converting the lateral offset feature HTi and the longitudinal offset feature ZTi into the calibration angle of the detection object, and adjusting the detection object according to the calibration angle.

8. The fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 7, characterized in that: The process of calibrating the angle adjustment of the detection object also includes: after the adjustment is completed, marking the captured image of the detection object as the adjustment image, obtaining the offset coefficient PY of the adjustment image and determining whether there is an offset abnormality in the adjustment image: if not, the calibration is completed; if so, regenerating the offset sequence and calibration angle according to the coordinates of the marked object i in the adjustment image, performing a second adjustment on the detection object using the calibration angle, and so on, until there is no offset abnormality in the detection object.

9. The fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 8, characterized in that: The specific process of the habit analysis module analyzing the offset habits of the sports venue video surveillance system includes: generating an analysis cycle, marking the lateral offset feature HTi and the longitudinal offset feature ZTi corresponding to the first calibration angle generated when the detection object is subjected to offset calibration analysis within the analysis cycle as lateral habit values HX and longitudinal habit values ZX respectively, performing variance calculation on all lateral habit values HX within the analysis cycle to obtain a lateral habit coefficient, and performing variance calculation on all longitudinal habit values ZX within the analysis cycle to obtain a longitudinal habit coefficient; and judging whether the detection object has lateral offset habits and longitudinal offset habits based on the lateral habit coefficient and the longitudinal habit coefficient.

10. The fault self-diagnosis and recovery system for a sports venue video surveillance system according to claim 9, characterized in that: The specific process of determining whether the detection object has the habit of lateral and longitudinal displacement includes: obtaining the habit threshold through the database; comparing the lateral habit coefficient and the longitudinal habit coefficient with the habit threshold respectively; if the lateral habit coefficient is less than the habit threshold, it is determined that the detection object has the habit of lateral displacement, and a lateral displacement processing signal is generated and sent to the mobile phone terminal of the administrator; if the longitudinal habit coefficient is less than the habit threshold, it is determined that the detection object has the habit of longitudinal displacement, and a longitudinal displacement processing signal is generated and sent to the mobile phone terminal of the administrator; otherwise, it is determined that the detection object does not have the habit of displacement.

Images