Camera fault detection and recovery method and device based on multiple sensors

Abstract

The invention relates to a camera fault detection and recovery method based on multiple sensors, and the method comprises the steps: S1, setting a preset type of sensors in a camera, and enabling each type of sensors to comprise a plurality of sensors for multi-source data collection; s2, acquiring multi-source data acquired by each type of sensors, detecting abnormal data in the multi-source data, and replacing the abnormal data with correction data based on machine learning; s3, counting an average value of each type of sensors, comparing the average value with a preset threshold value, and calculating a fault probability value according to a comparison result; s4, based on the evidence theory, determining the fault type and position of the camera according to the fault probability value corresponding to each type of sensor; and S5, according to the fault type and position, executing recovery processing or alarming.

Description

Technical Field

[0001] The present invention relates to the technical field of video surveillance, and in particular to a method and device for detecting and recovering camera faults based on multiple sensors. Background Art

[0002] Security monitoring equipment such as cameras play a key role in commanding the construction of cities. As intelligent sensing terminals, cameras are widely used in many fields such as urban management, security prevention, traffic monitoring, and environmental monitoring through video image acquisition, data analysis, and intelligent processing.

[0003] Cameras may face a variety of faults in complex working environments. Cameras need to be regularly inspected and maintained during operation. Camera fault detection in traditional technology mainly judges the quality of video images, ignoring the impact of environmental factors on its hardware, which may affect the accuracy of fault detection. In addition, after a fault occurs, there is a lack of fast and effective recovery methods, and the camera monitoring interruption time is long, affecting the security effect. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to provide a camera fault detection and recovery method based on multiple sensors. By integrating various types of sensors, the camera hardware is monitored in different dimensions, and the data obtained from different sensors are integrated and analyzed through a data fusion algorithm. When abnormal data is monitored, the fault is identified and classified, and the location and type of the fault are determined.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present application provides a camera fault detection and recovery method based on multiple sensors, the method comprising:

[0007] S1, setting preset categories of sensors in the camera, each category of sensors including a plurality of sensors for multi-source data collection;

[0008] S2, acquiring multi-source data collected by each type of sensor, detecting abnormal data in the multi-source data, and replacing the abnormal data with corrected data based on machine learning;

[0009] S3, counting the average value of each type of sensor, and comparing it with the preset threshold, and calculating the fault probability value according to the comparison result;

[0010] S4, based on the evidence theory, according to the fault probability value corresponding to each type of sensor, determine the fault type and location of the camera;

[0011] S5: Perform recovery processing or alarm according to the fault type and location.

[0012] In one implementation, the preset category of sensors includes: temperature sensors, voltage / current sensors, vibration sensors, and acceleration sensors.

[0013] In an implementation manner, in S2, abnormal data in multi-source data is determined according to the Lagerbus criterion.

[0014] In one implementation, in S2, the input abnormal data is replaced based on the pre-trained CFA-PSO-RBF neural network.

[0015] In one implementation, the fault types include recoverable faults and unrecoverable faults; for recoverable faults, recovery processing is performed; for unrecoverable faults, alarm processing is performed.

[0016] In a second aspect, a multi-sensor based camera fault detection and recovery device is provided, comprising: a main power module, a backup power module, a main control module, a temperature sensor, a voltage / current sensor, a vibration sensor, an acceleration sensor, and an alarm module;

[0017] The main control module is used to obtain the multi-source data collected by each type of sensor, detect abnormal data in the multi-source data, and replace the abnormal data with corrected data based on machine learning; and to count the average value of each type of sensor, and compare it with the preset threshold, and calculate the fault probability value according to the comparison result; based on the evidence theory, according to the fault probability value corresponding to each type of sensor, determine the fault type and location of the camera; and according to the fault type and location, perform recovery processing or alarm.

[0018] The present invention adopts the above technical scheme, integrates multiple sensors, and adopts advanced data fusion algorithms to realize automatic recovery and remote monitoring and diagnosis functions, thereby effectively solving the problems of insufficient accuracy and timeliness, high false alarm rate and missed alarm rate, and slow recovery speed in traditional camera fault detection and recovery, and is of great significance in ensuring the stable operation of the camera and improving the reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Schematic diagram of the hardware structure of a multi-sensor based camera fault detection and recovery device according to an embodiment of the present invention;

[0020] Figure 2 Schematic diagram of the network architecture of the algorithm according to the embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0022] In view of the above problems in the prior art, an embodiment of the present invention provides a camera fault detection and recovery method based on multiple sensors, the method comprising:

[0023] S1, setting preset categories of sensors in the camera, each category of sensors including a plurality of sensors for multi-source data collection;

[0024] S2, acquiring multi-source data collected by each type of sensor, detecting abnormal data in the multi-source data, and replacing the abnormal data with corrected data based on machine learning;

[0025] S3, counting the average value of each type of sensor, and comparing it with the preset threshold, and calculating the fault probability value according to the comparison result;

[0026] S4, based on the evidence theory, according to the fault probability value corresponding to each type of sensor, determine the fault type and location of the camera;

[0027] S5: Perform recovery processing or alarm according to the fault type and location.

[0028] Accordingly, an embodiment of the present invention provides a multi-sensor based camera fault detection and recovery device, comprising: a main power module, a backup power module, a main control module, a temperature sensor, a voltage / current sensor, a vibration sensor, an acceleration sensor, and an alarm module;

[0029] The main control module is used to obtain the multi-source data collected by each type of sensor, detect abnormal data in the multi-source data, and replace the abnormal data with corrected data based on machine learning; and to count the average value of each type of sensor, and compare it with the preset threshold, and calculate the fault probability value according to the comparison result; based on the evidence theory, according to the fault probability value corresponding to each type of sensor, determine the fault type and location of the camera; and according to the fault type and location, perform recovery processing or alarm.

[0030] The above method and device are described in a more detailed embodiment below in conjunction with more drawings of the present application.

[0031] The hardware system of the device of the present invention is composed of a main power module, a backup power module (battery pack), a main control module, a temperature sensor, a voltage / current sensor, a vibration sensor, an acceleration sensor, an alarm module, etc. Figure 1 shown.

[0032] In this example, the main control module adopts the RK3399 module.

[0033] The main power module provides a stable operating voltage to the RK3399 control module and the image acquisition module.

[0034] The temperature sensor mainly detects the temperature of the fill light board and the internal temperature of the device, sets a threshold, and when the detected temperature exceeds the threshold, the RK3399 control module controls the fan for heat dissipation.

[0035] The vibration sensor mainly detects camera failures caused by loose components inside the camera. Such failures are generally non-automatic recovery failures. After such failure signals are uploaded to the RK3399 control module for processing, the drive alarm module reminds relevant personnel to conduct inspections.

[0036] The acceleration sensor mainly detects illegal movement or vandalism of the camera. Once an abnormality is detected, the data is uploaded to the RK3399 control module for processing and analysis, triggering the alarm mechanism to notify the monitoring center or relevant personnel for processing.

[0037] The voltage / current sensor monitors the power supply of the camera, including the stability and fluctuation range of current and voltage.

[0038] The alarm module is mainly used to inform the monitoring center or relevant personnel to handle the problem when an irrecoverable fault occurs in the camera.

[0039] The backup power supply (battery pack) is used to start the backup power supply when there is a problem with the main power supply inside the device, so that the camera can work normally and continuously.

[0040] In the embodiment of the present invention, when collecting data at the sensor front end, an advanced multi-source data fusion algorithm (such as CFA-PSO-RB, DS algorithm) is used to integrate and analyze data from different sensors, monitor the state of the speed camera in all directions, improve the accuracy of fault detection, and enhance the processing of weak signals to improve the sensitivity of fault detection.

[0041] In camera fault detection, the data collected by the front-end temperature sensor, voltage / current sensor, vibration sensor, and acceleration sensor need to be converted first because of the different characteristics of the data collected, so that the system can identify the collected data signal, and the data collected by each sensor is converted into an electrical signal through A / D conversion. Each sensor is susceptible to external interference, and the collected data may be affected, so the abnormal data needs to be corrected. In order to ensure that the data collected by each sensor is true and reliable, the CFA-PSO-RBF neural network algorithm based on the Lagerbus criterion is adopted in the present invention. First, the residual value of the collected temperature, vibration parameters, voltage / current, and acceleration sensor parameters is calculated using the Lagerbus criterion. If the residual value is large, the data collected by the sensor is abnormal data, and then the neural network prediction model is trained with normal data (using the CFA-PSO-RBF neural network prediction model to predict the abnormal data), and the abnormal data is replaced. Finally, the arithmetic mean algorithm is used to fuse and analyze the data collected by various sensors of the same type, and compare them with the threshold value set initially. After the fusion of the same type of sensor data is completed, BPAF is used to assign fault probability values ​​to the fusion results of the same type of data. Then, the DS algorithm is used to fuse the data collected by different types of sensors to obtain the results and clarify the fault type and location of the camera. The algorithms involved in the device have adaptive learning capabilities and can automatically adjust the corresponding parameters. Multi-source data fusion network model Figure 2 shown.

[0042] To summarize, this application effectively solves the problems of insufficient accuracy and timeliness, high false alarm and missed alarm rates, and slow recovery speed in traditional camera fault detection and recovery by integrating multiple sensors and adopting advanced data fusion algorithms to achieve automatic recovery and remote monitoring and diagnosis functions. It is of great significance in ensuring the stable operation of the camera and improving the reliability of the system.

[0043] In the several embodiments provided by the present invention, it should be understood that the disclosed method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0044] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (Processor) to perform some steps of the above-mentioned method of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program codes.

[0045] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A camera fault detection and recovery method based on multiple sensors, characterized in that: The method comprises: S1, setting preset categories of sensors in the camera, each category of sensors including a plurality of sensors for multi-source data collection; S2, acquiring multi-source data collected by each type of sensor, detecting abnormal data in the multi-source data, and replacing the abnormal data with corrected data based on machine learning; S3, counting the average value of each type of sensor, and comparing it with the preset threshold, and calculating the fault probability value according to the comparison result; S4, based on the evidence theory, according to the fault probability value corresponding to each type of sensor, determine the fault type and location of the camera; S5: Perform recovery processing or alarm according to the fault type and location.

2. The multi-sensor based camera fault detection and recovery method according to claim 1, characterized in that: The preset categories of sensors include: temperature sensors, voltage / current sensors, vibration sensors, and acceleration sensors.

3. The multi-sensor based camera fault detection and recovery method according to claim 1, characterized in that: In S2, abnormal data in the multi-source data is determined according to the Lagerbus criterion.

4. The multi-sensor based camera fault detection and recovery method according to claim 3, characterized in that: In S2, the input abnormal data is replaced based on the pre-trained CFA-PSO-RBF neural network.

5. The multi-sensor based camera fault detection and recovery method according to claim 1, characterized in that: The fault types include recoverable faults and unrecoverable faults; for recoverable faults, recovery processing is performed; For irrecoverable faults, perform alarm processing.

6. A camera fault detection and recovery device based on multiple sensors, characterized in that: include: Main power module, backup power module, main control module, temperature sensor, voltage / current sensor, vibration sensor, acceleration sensor, alarm module; A main control module, used to obtain multi-source data collected by each type of sensor, detect abnormal data in the multi-source data, and replace the abnormal data with corrected data based on machine learning; The average value of each type of sensor is counted and compared with the preset threshold, and the fault probability value is calculated based on the comparison result; based on the evidence theory, the fault type and location of the camera are determined according to the fault probability value corresponding to each type of sensor; And according to the fault type and location, perform recovery processing or alarm.

Images