Security and protection device for monitoring image real-time analysis and abnormity early warning

Abstract

The invention discloses a monitoring image real-time analysis and abnormity early warning security and protection device, which comprises a camera, an audible and visual alarm is arranged at a lens of the camera, a data acquisition module, a preprocessing module, a network transmission module, an image analysis module and an abnormity early warning module are integrated in the camera, and a memory is arranged at the bottom of the camera. The memory is internally provided with a data caching module, the data caching module is internally provided with a caching unit with an intelligent exception handling and recovery mechanism, the data acquisition module acquires images, the data preprocessing module compresses and encodes the images to reduce the data volume, the network transmission module transmits data by using double network interfaces and flow control, and when network delay occurs, the data can be transmitted to the data acquisition module. The data is stored in the data caching module, after the network is normal, the data is sent to the image analysis module, and the abnormity early warning module triggers the audible and visual alarm to give an early warning according to the result. According to the method, an intelligent exception handling and recovery mechanism is adopted, when network delay happens to the data, continuous operation is still carried out, and the risk of data loss is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of security technology, and specifically relates to a security device for real-time analysis of surveillance images and abnormal warning. Background Art

[0002] At present, with the continuous development of the economy, people have put forward higher requirements for the performance of security systems. Surveillance devices, such as common surveillance cameras, as key components of security systems, bear the important task of directly capturing images, just like the "visual center" of the entire security system.

[0003] In the actual operation process of existing security devices for real-time analysis of surveillance images and abnormal warning, when network transmission is delayed, due to the lack of an effective data caching mechanism, it is unable to continuously store surveillance image data. Once the network fluctuates, or when the cache is full, data transmission is prone to interruption, resulting in situations such as stuttering of surveillance images and data loss, and it is unable to provide complete data support for subsequent image analysis. Summary of the Invention

[0004] The purpose of the present invention is to provide a security device for real-time analysis of surveillance images and abnormal warning in order to solve the technical problems that when the network is delayed, the existing device freezes or even stops working due to poor data transmission, affecting the real-time performance of security surveillance and causing loss of surveillance image data.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A security device for real-time analysis of surveillance images and abnormal warning, comprising:

[0007] A camera, on which an acoustic-optic alarm is fixedly installed at the lens of the camera, and a data acquisition module, a data preprocessing module, a network transmission module, an image analysis module, and an abnormal warning module are provided inside the camera;

[0008] A memory, installed at the bottom of the camera, a data caching module is provided inside the memory, a caching unit is provided inside the data caching module, and an intelligent abnormal processing and recovery mechanism is provided inside the caching unit;

[0009] The data acquisition module acquires surveillance images, compresses and encodes them through the data preprocessing module to reduce the data volume, and then the network transmission module transmits the data through a dual network interface and flow control. When there is a delay, the data is stored in the caching module inside the memory, and the intelligent abnormal processing and recovery mechanism inside its intelligent caching unit ensures data management. After the network is normal, the data is sent to the image analysis module for processing, and finally the abnormal warning module triggers the acoustic-optic alarm for warning according to the analysis result.

[0010] Among them, the intelligent exception handling and recovery mechanism includes data reception, cache space check, data priority judgment and discard, data writing, network status monitoring, data reading and transmission, cache cleaning and update, and exception handling. The data reception link can continuously receive pre-processed image data. After the cache space check, the remaining cache capacity is checked before each reception. If the cache space is sufficient, it directly enters the data writing. If the cache space is insufficient, through data priority judgment and discard, the priority is evaluated based on time, analysis status, etc., and low-priority data is discarded. After confirming that the cache space is sufficient, it enters the data writing. Then, the network status monitoring continuously monitors the network latency and packet loss rate. Through data reading and transmission, when the network is normal or the analysis module needs it, it reads and transmits in order to the subsequent module. Then, after cache cleaning and update, the space is released after the data is successfully transmitted, and it is optimized regularly. Finally, when a read / write error exception occurs, exception handling is performed, recorded and repaired, and reported to the system at the same time.

[0011] Among them, in the data priority judgment and discard, the cache module will establish a priority evaluation system, divide the data according to the reception time, and newly received data usually has a higher priority. Then, unanalyzed data and abnormal data have a higher priority than analyzed data, and the monitoring data in key areas has a higher priority than the monitoring data in ordinary areas.

[0012] Among them, in the exception handling process, when an abnormal situation such as data writing failure, reading error, cache overflow, or network interruption is detected, the exception handling process can be immediately triggered. At the same time, the cache module will record in detail the time, type, data block number involved, and relevant operation step information when the exception occurs. Then, according to different exception types, corresponding repair measures are taken. When it is detected that the data transmission is blocked due to a network exception, on the one hand, the data writing operation is suspended to prevent further cache backlog, and on the other hand, continuous attempts are made to reconnect to the network, with a connection attempt every [X] seconds. After the network is restored, the data transmission is automatically resumed, and the data that was not completed due to the network interruption is retransmitted. During the exception handling process, the cache module will send an exception report to the system control module of the security device in a timely manner, including the detailed situation of the exception, the handling measures taken, and the handling results. When the abnormal situation still cannot be solved after multiple automatic processes, the cache module will prompt the need for manual intervention by means of an audible and visual alarm.

[0013] Among them, the data preprocessing module includes an image compression unit and a coding conversion unit. The image compression unit receives image data from the camera and converts it into a coding format suitable for network transmission and subsequent processing. The image compression unit deeply compresses the coded image, dynamically adjusts the compression ratio according to the complexity of the image content. For the background area with less details, a higher compression ratio can be adopted. For the area where the key target is located, the compression ratio is appropriately reduced to retain more information. At the same time, the hardware acceleration technology is used to quickly complete the compression operation.

[0014] Among them, the network transmission module has dual network interfaces, including an Ethernet interface and a wireless network card interface, and can automatically switch network connections according to the network conditions. The network transmission module is also provided with a network traffic control unit, and the network traffic control unit uses the token bucket algorithm to control the network traffic and gives priority to ensuring the transmission of image data.

[0015] Among them, the image analysis module uses a multi-core processor and an AI acceleration chip to work together. The multi-core processor is used to process image data in parallel, and the AI acceleration chip is used to execute deep learning algorithms for image recognition and analysis.

[0016] Among them, when triggering the sound and light alarm, the abnormal warning module can also send abnormal information to the preset receiving terminal by means of text messages and emails.

[0017] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are:

[0018] In the present invention, due to the setting of the intelligent exception handling and recovery mechanism in the cache unit, when a data write failure occurs, the intelligent exception handling and recovery mechanism will automatically retry and adjust the write parameters, or during a network interruption, the data will be temporarily stored in the cache and continue to be transmitted after the network is restored, ensuring that the data will not be lost due to exceptions. At the same time, when the cache is nearly full, the intelligent exception handling and recovery mechanism will start the data priority judgment and discard strategy, actively release the cache space, and preferentially discard the low-priority data to ensure that the cache space can provide storage for the high-priority data and avoid the loss of important data due to cache overflow, optimizing the use efficiency of the cache. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a schematic diagram of the overall structure of the present invention;

[0020] Figure 2 is a flowchart of the operation of the internal module in the present invention;

[0021] Figure 3 is a flowchart of the operation of the intelligent exception handling and recovery mechanism in the present invention.

[0022] Markings in the figure: 10, camera; 11, acoustic-optic alarm; 20, memory. Specific implementation manner

[0023] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0024] Refer to Figures 1-3 , a security device for real-time analysis of surveillance images and abnormal early warning, including a camera 10. An acoustic-optic alarm 11 is fixedly installed at the lens of the camera 10. A data acquisition module, a data preprocessing module, a network transmission module, an image analysis module and an abnormal early warning module are provided in the camera 10. The bottom of the camera 10 is fixedly connected with a memory 20. A data cache module is provided in the memory 20. A cache unit is provided in the data cache module. An intelligent abnormal processing and recovery mechanism is provided in the cache unit. The data acquisition module acquires surveillance images, compresses and encodes them through the data preprocessing module to reduce the data volume, and then the network transmission module transmits the data through a dual network interface and traffic control. In case of delayed data, it is stored in the cache module in the memory 20. The intelligent cache unit's internal intelligent abnormal processing and recovery mechanism ensures data management. After the network is normal, the data is sent to the image analysis module for processing. Finally, the abnormal early warning module triggers the acoustic-optic alarm 11 for early warning according to the analysis result.

[0025] Specifically, during the process of network transmission, the problem of network delay is likely to occur. The added memory 20 and the data cache module therein can serve as a temporary storage area during network delay to receive and store the data waiting to be transmitted after being processed by the data preprocessing module, avoiding data transmission interruption or loss caused by network delay, ensuring the continuity of data in the transmission link, and guaranteeing the smooth data flow of the entire security device. The intelligent abnormal processing and recovery mechanism in the cache unit can monitor and process possible errors in real time during the data reading and writing process. When the data writing to the cache fails, the mechanism can automatically attempt to rewrite and adjust the writing parameters. If an error occurs during data reading, it can use a verification algorithm for repair or recover from backup data, which ensures the integrity of the data stored in the cache, provides an accurate and reliable data basis for subsequent image analysis. At the same time, when the data cache is full, a data priority judgment and discard strategy will be activated to actively release the cache space, preferentially discard low-priority data, and ensure that the cache space can provide storage for high-priority data, avoiding the loss of important data caused by cache overflow.

[0026] Refer to Figure 3As shown, the intelligent exception handling and recovery mechanism includes data reception, cache space check, data priority judgment and discard, data writing, network condition monitoring, data reading and transmission, cache cleaning and update, and exception handling. In the data reception link, the pre-processed image data can be continuously received. Regardless of the network condition, the data cache module always maintains the state of receiving data, providing a basis for subsequent processing. After the cache space check, the remaining cache capacity is checked before each reception. If the cache space is sufficient, it directly enters the data writing process. If the cache space is insufficient, through data priority judgment and discard, the priority is evaluated based on time, analysis status, etc., and the low-priority data is discarded. After confirming that the cache space is sufficient, it enters the data writing process. The newly received image data is written into the relatively idle storage area in the cache according to the first-in, first-out principle. The cache management mechanism will record the storage location and related information of each data block for convenient subsequent reading and management. Then, the network condition monitoring module continuously monitors the network latency and packet loss rate. Once it is found that the network has a delay or instability, the module will timely adjust the working mode to ensure that the data will not be lost or damaged due to network problems. Through data reading and transmission, when the network is normal or the analysis module needs it, the data is read in sequence and transmitted to the subsequent module, and the data stored earliest is preferentially read to ensure the continuity and timeliness of data processing. Then, after cache cleaning and update, the space is released after the data is successfully transmitted and optimized regularly. Finally, when a read / write error exception occurs, exception handling is performed, recorded and repaired, and reported to the system at the same time for fault troubleshooting and repair.

[0027] Refer to Figure 3 As shown, in data priority judgment and discard, the cache module will establish a priority evaluation system, dividing the data according to the reception time. Newly received data usually has a higher priority because new data can better reflect the current real-time situation, which is crucial for timely detecting abnormalities. Then, unanalyzed data and abnormal data have a higher priority than analyzed data. Since unanalyzed data may hide undetected abnormal situations, giving priority to retaining it can ensure that the security device does not miss important information. If there are suspected abnormal features in the image, such as abnormal gathering of people, abnormal movement of items, etc., the corresponding data will be given a higher priority. The importance of different monitoring areas is different, and the priority of the data generated by them is also different. The monitoring data of key areas has a higher priority than that of ordinary areas because abnormal situations in key areas may have a significant impact on production safety and enterprise operations.

[0028] Refer to Figure 3As shown, in the exception handling process, when detecting exceptions such as data write failure, read error, cache overflow, or network interruption, the exception handling process can be immediately triggered. During the data writing process, if consecutive write operations all return error codes, it is determined as a write failure exception. Meanwhile, the cache module will record in detail the time, type, data block number involved, and relevant operation step information when the exception occurs. These records not only help to troubleshoot the root cause of problems later but also provide a basis for system fault diagnosis. Then, according to different exception types, corresponding repair measures are taken. If the data write fails, it may attempt to rewrite the operation, with a maximum of [X] retries. Before each retry, the write parameters will be appropriately adjusted. If there is a read error, the data verification algorithm will be used to verify and repair the data. If the repair fails, it will attempt to recover from the backup data. For the cache overflow exception, the data priority judgment and discard process will be immediately started to release the cache space, and at the same time, a cache emergency status notification will be sent to the system control module. When it is detected that network anomalies cause data transmission to be blocked, on the one hand, the data write operation will be suspended to prevent further cache backlog, and on the other hand, continuous attempts will be made to reconnect to the network, with a connection attempt every [X] seconds. After the network is restored, the data transmission will be automatically resumed, and the data that was not completely transmitted due to the network interruption will be retransmitted. During the exception handling process, the cache module will promptly send an exception report to the system control module of the security device, including the detailed situation of the exception, the handling measures taken, and the handling results. When the exception still cannot be resolved after multiple automatic processes, the cache module will prompt the need for manual intervention by means of the sound and light alarm 11, and at the same time, detailed exception information will be displayed on the system interface to facilitate technicians to troubleshoot and repair faults.

[0029] Refer to Figure 2 As shown, the data preprocessing module includes an image compression unit and a coding conversion unit. The image compression unit receives the image data from the camera 10 and converts it into a coding format suitable for network transmission and subsequent processing. The image compression unit deeply compresses the coded image and dynamically adjusts the compression ratio according to the complexity of the image content. For the background area with less detail, a higher compression ratio can be used. For the area where the key target is located, the compression ratio is appropriately reduced to retain more information. At the same time, the hardware acceleration technology is used to quickly complete the compression operation. The image data is converted into a coding format suitable for network transmission by the coding conversion unit and then deeply compressed by the image compression unit, significantly reducing the volume of the image data. During network transmission, a smaller data volume means a lower network bandwidth requirement, reducing the burden of network transmission. Especially in an unstable network environment or with limited bandwidth, it can effectively avoid problems such as transmission delay, lag, and even data loss caused by excessive data volume, ensuring that the image data can be smoothly transmitted to the subsequent processing module.

[0030] Refer toFigure 2 As shown, the network transmission module has dual network interfaces, including an Ethernet interface and a wireless network card interface, and can automatically switch network connections according to the network conditions. The network transmission module is also equipped with a network traffic control unit. The network traffic control unit uses the token bucket algorithm to control network traffic, giving priority to ensuring the transmission of image data. The token bucket algorithm can effectively control the data sending rate and avoid network congestion caused by too fast data sending. When the network traffic is large, the algorithm will reasonably limit the data sending volume according to the generation speed and quantity of tokens, keeping the network traffic at a stable level. This can not only ensure the transmission quality of image data, but also improve the stability and reliability of the entire network, and avoid affecting other devices or services due to network congestion.

[0031] Referring to Figure 2 As shown, the image analysis module uses a multi-core processor and an AI acceleration chip to work together. The multi-core processor is used to process image data in parallel, and the AI acceleration chip is used to execute deep learning algorithms for image recognition and analysis. In the image analysis module, the multi-core processor processes image data in parallel, can quickly split tasks and perform operations simultaneously, process a large number of high-resolution images in a short time, improve the processing speed, and ensure the real-time monitoring. The AI acceleration chip focuses on executing deep learning algorithms, accurately identifying target objects and behavior patterns in complex scenarios, enhancing the analysis ability. The two work together in division of labor, optimize resource utilization, reduce energy consumption, and also enable the security device to intelligently analyze and give early warnings, detect abnormalities in a timely manner, and adapt to demand upgrades, improving the security intelligence level.

[0032] Referring to Figure 2 As shown, when triggering the sound and light alarm 11, the abnormal warning module can also send abnormal information to the preset receiving terminals by means of text messages and emails. The multi-channel notifications are backed up with each other to prevent the failure of a single notification method. If the sound and light alarm 11 is not noticed due to a noisy on-site environment, or the text message sending is affected by a network failure, the email may still be successfully delivered, ensuring that the abnormal information can be conveyed, and guaranteeing that the security system can give effective early warnings in various complex situations. As written notification forms, text messages and emails can retain the abnormal information for subsequent traceability.

[0033] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A security device for real-time analysis of monitoring images and abnormal warning, characterized in that: include: A camera (10), wherein a sound and light alarm (11) is fixedly mounted at the lens of the camera (10), and the camera (10) is provided with a data acquisition module, a data preprocessing module, a network transmission module, an image analysis module and an abnormal warning module; A memory (20) is installed at the bottom of the camera (10), wherein a data cache module is provided in the memory (20), a cache unit is provided in the data cache module, and an intelligent exception processing and recovery mechanism is provided in the cache unit; The data acquisition module acquires monitoring images, which are compressed and encoded by the data preprocessing module to reduce the data volume. The network transmission module then transmits the data by means of dual network interfaces and flow control. In case of delay, the data is stored in a cache module in the memory (20). The intelligent abnormality processing and recovery mechanism inside the intelligent cache unit ensures data management. After the network is normal, the data is sent to the image analysis module for processing. Finally, the abnormality warning module triggers the sound and light alarm (11) to issue a warning according to the analysis results.

2. The security device for real-time analysis of monitoring images and abnormal warning according to claim 1, characterized in that: The intelligent exception handling and recovery mechanism includes data reception, cache space check, data priority judgment and abandonment, data writing, network status monitoring, data reading and transmission, cache cleaning and updating, and exception handling. The data receiving link can continuously receive pre-processed image data, and after cache space check, check the remaining cache capacity before each reception. If the cache space is sufficient, directly enter the data writing. If the cache space is insufficient, through data priority judgment and abandonment, evaluate the priority based on time, analysis status, etc., abandon low-priority data, and enter data writing after confirming that the cache space is sufficient. Then, the network status monitoring is used to grasp the network delay and packet loss rate in real time. Through data reading and transmission, when the network is normal or the analysis module needs it, it is read in sequence and transmitted to the subsequent module. Then, after cache cleaning and updating, the space is released after the data is successfully transmitted, and regular optimization is performed. Finally, when read and write errors occur, exception handling is performed, recorded and repaired, and reported to the system at the same time.

3. The security device for real-time analysis of monitoring images and abnormal warning according to claim 2, characterized in that: The cache module in the data priority judgment and discarding will establish a priority evaluation system to divide the data according to the receiving time. Newly received data usually has a higher priority, and then unanalyzed data and abnormal data have a higher priority than analyzed data, and monitoring data in key areas have a higher priority than monitoring data in ordinary areas.

4. The security device for real-time analysis of monitoring images and abnormal warning according to claim 2, characterized in that: In the exception handling process, when an abnormal situation such as data writing failure, reading error, cache overflow or network interruption is detected, the exception handling process can be triggered immediately. At the same time, the cache module will record in detail the time, type, data block number involved and related operation step information of the abnormality, and then take corresponding repair measures according to different abnormality types. When it is detected that the network abnormality causes data transmission to be blocked, on the one hand, the data writing operation is suspended to prevent the cache from further backlog, and on the other hand, attempts are continuously made to reconnect to the network, and a connection attempt is made every [X] seconds. After the network is restored, data transmission is automatically resumed, and data that has not been transmitted due to network interruption is reissued. During the exception handling process, the cache module will promptly send an exception report to the system control module of the security device, including the details of the abnormality, the processing measures taken and the processing results. If the abnormal situation cannot be resolved after multiple automatic processing, the cache module will prompt the need for manual intervention through an audible and visual alarm (11).

5. The security device for real-time analysis of monitoring images and abnormal warning according to claim 1, characterized in that: The data preprocessing module comprises an image compression unit and a coding conversion unit. The image compression unit receives image data from a camera (10) and converts the image data into a coding format suitable for network transmission and subsequent processing. The image compression unit performs deep compression on the coded image and dynamically adjusts the compression ratio according to the complexity of the image content. For background areas with fewer details, a higher compression ratio can be used. For areas where key targets are located, the compression ratio is appropriately reduced to retain more information. At the same time, the compression operation is quickly completed using hardware acceleration technology.

6. The security device for real-time analysis of monitoring images and abnormal warning according to claim 1, characterized in that: The network transmission module has dual network interfaces, including an Ethernet interface and a wireless network card interface, and can automatically switch network connections according to network conditions. The network transmission module is also provided with a network traffic control unit, which uses a token bucket algorithm to control network traffic and give priority to ensuring the transmission of image data.

7. The security device for real-time analysis of monitoring images and abnormal warning according to claim 1, characterized in that: The image analysis module uses a multi-core processor and an AI acceleration chip to work together. The multi-core processor is used to process image data in parallel, and the AI ​​acceleration chip is used to execute a deep learning algorithm for image recognition and analysis.

8. The security device for real-time analysis of monitoring images and abnormal warning according to claim 1, characterized in that: The abnormal warning module can also send abnormal information to a preset receiving terminal by means of text messages or emails while triggering the sound and light alarm (11).

Images