Real-time video behavior identification and abnormity monitoring system and method based on edge calculation

Abstract

The invention relates to the technical field of smart home, and discloses a real-time video behavior recognition and anomaly monitoring system based on edge computing, which comprises an edge computing node, a video preprocessing module, a background separation and target detection module, an alarm triggering and equipment linkage module, a behavior recognition module and an anomaly monitoring module, the edge computing node is connected with the video preprocessing module and the alarm triggering and equipment linkage module, and the edge computing node is connected with the video preprocessing module and the alarm triggering and equipment linkage module. The system has multiple advantages of real-time performance and high efficiency, privacy protection and security, accurate identification and low false alarm rate, equipment linkage and intelligence, strong adaptability and expandability, energy conservation and environmental protection, and the like, and the system has wide application prospects and important practical values in the field of smart home security protection due to the advantages.

Description

Technical Field

[0001] The present invention relates to the field of smart home technology, and in particular to a real-time video behavior recognition and anomaly monitoring system and method based on edge computing. Background Art

[0002] With the advancement of technology and the popularization of the Internet, smart home systems have become an important part of modern family life. Especially in the field of home security monitoring, video surveillance systems have been widely installed and applied due to their intuitive and real-time characteristics. However, the current video surveillance solutions mainly rely on remote cloud servers for data storage and in-depth analysis, and the following problems have gradually been exposed in practical applications:

[0003] 1. Since video data needs to be uploaded to a remote server for analysis and processing, the delay in data transmission leads to a slow response speed of the monitoring system, making it impossible to provide timely feedback and processing of real-time events;

[0004] 2. Large network bandwidth consumption: The amount of video data is large, and uploading it to the cloud requires a lot of network bandwidth, which puts pressure on the home network environment. Especially in an environment with poor network conditions, video transmission is unstable, affecting the monitoring effect;

[0005] 3. When video data is transmitted and stored in the network, there is a risk of being intercepted and leaked, and the user's privacy and security cannot be effectively protected. Therefore, we propose a real-time video behavior recognition and anomaly monitoring system and method based on edge computing to solve the above problems. Summary of the invention

[0006] The present invention provides a real-time video behavior recognition and anomaly monitoring system and method based on edge computing, which can effectively solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a real-time video behavior recognition and anomaly monitoring system based on edge computing, including an edge computing node, a video preprocessing module, a background separation and target detection module, an alarm triggering and equipment linkage module, a behavior recognition module and an anomaly monitoring module, the edge computing node is connected to the video preprocessing module and the alarm triggering and equipment linkage module respectively, the video preprocessing module is connected to the background separation and target detection module, the background separation and target detection module is connected to the behavior recognition module, the behavior recognition module is connected to the anomaly monitoring module, and the anomaly monitoring module is connected to the alarm triggering and equipment linkage module.

[0008] Preferably, the edge computing node is integrated into a smart home camera or sensor device, and is equipped with a high-performance processor and a dedicated algorithm for local analysis of real-time video streams. The edge computing node uses edge computing technology to achieve local processing and analysis of video data, thereby reducing data transmission delays and bandwidth consumption.

[0009] Preferably, the video preprocessing module performs denoising and format conversion preprocessing operations on the video data to ensure the quality and consistency of the video data. The video preprocessing module improves the clarity and stability of the video data through image processing technology, providing a basis for subsequent behavior recognition and abnormality monitoring.

[0010] Preferably, the background separation and target detection module adopts advanced background separation technology to identify moving targets in the video and accurately detect them. The background separation and target detection module uses image processing technology such as background difference method or optical flow method to identify moving targets in the foreground and accurately detect them by comparing the difference between the current video frame and the background model.

[0011] Preferably, the behavior recognition module identifies the detected target behavior based on machine learning or deep learning algorithms, including specific behavior patterns of people, animals, and vehicles. The behavior recognition module analyzes the target's motion trajectory, speed, and posture characteristics, and uses a pre-trained model or algorithm to identify the target behavior, such as walking, running, and staying.

[0012] Preferably, machine learning is an artificial intelligence technique that, given a data set, can automatically learn patterns and use the learning results to make predictions or decisions. It relies on mathematical and statistical algorithms to build models that allow computers to learn autonomously without being explicitly programmed;

[0013] Deep learning is a machine learning method that uses a neural network composed of multiple levels to simulate and solve complex problems. Its core idea is to gradually improve the model's ability to represent and predict data through hierarchical learning and feature extraction.

[0014] Preferably, the anomaly monitoring module monitors abnormal behaviors or events in the video according to preset rules or models, and immediately triggers an alarm once an anomaly is detected. The anomaly monitoring module can identify abnormal patterns that are inconsistent with the target behavior, such as intrusion, fall, and lost items, and trigger corresponding alarms or notifications.

[0015] Preferably, the alarm triggering and device linkage module is linked with other devices in the smart home system to form a complete intelligent security system. The alarm triggering and device linkage module is connected and communicated with other devices through the communication protocol or interface of the smart home system to realize the alarm triggering and device linkage functions.

[0016] The present invention also provides a real-time video behavior recognition and anomaly monitoring method based on edge computing, comprising the following steps:

[0017] S1, initialization operation, when the edge computing node is started, the necessary software modules are loaded, the network connection is configured, the hardware status is checked, and the algorithms and models required for video processing are prepared, and then the connection status is checked to ensure that the video stream can be stably transmitted to the edge computing node, and then the alarm triggering conditions are set, the device linkage rules are configured, and the recognition criteria for abnormal behavior are defined;

[0018] S2, the edge computing node receives the video stream from the smart home camera in real time through a wireless network or a wired connection, monitors the transmission quality of the video stream, ensures that the image is clear, without freezes or delays, and temporarily stores the received video stream in the buffer of the edge computing node;

[0019] S3, the video preprocessing module preprocesses the video stream, including denoising, format conversion and image enhancement;

[0020] S4, the background separation and target detection module identifies the moving target in the foreground by comparing the difference between the current video frame and the background model, and analyzes the foreground image, identifies the specific moving target, tracks the moving trajectory of the target, and records the position, speed and direction information;

[0021] S5. The behavior recognition module uses machine learning or deep learning algorithms to perform pattern analysis on the target behavior, then extracts key features from the target's motion trajectory and posture changes, matches the extracted features with known behavior patterns, determines the target's current behavior type, and then evaluates the normality or abnormality of the behavior, providing a basis for subsequent abnormality monitoring;

[0022] S6, the abnormal monitoring module detects abnormal behaviors or events in the video according to preset rules or models;

[0023] S7. Once abnormal behavior is detected to trigger an alarm, the user or relevant security personnel will be notified. At the same time, it can be linked with other smart home devices, such as turning on lights, starting camera recording, and sending notifications to mobile phone apps. During this process, relevant information of abnormal events can be recorded, including timestamps, locations, target types, behavior types, and triggered alarms and linkage responses;

[0024] S8. The system will record relevant information of abnormal events. According to the results of abnormal events, the system may update the background model to adapt to environmental changes or improve the accuracy of recognition;

[0025] S9. If no abnormal behavior is detected, continue to monitor the video stream and repeat the process from S1 to S8 to collect new video data and update the internal state.

[0026] Beneficial Effects

[0027] The present invention provides a real-time video behavior recognition and anomaly monitoring system and method based on edge computing. Compared with the prior art, it has the following beneficial effects:

[0028] This real-time video behavior recognition and anomaly monitoring system based on edge computing has multiple advantages such as real-time and high efficiency, privacy protection and security, accurate recognition and low false alarm rate, equipment linkage and intelligence, strong adaptability and scalability, energy saving and environmental protection. These advantages make the system have broad application prospects and important practical value in the field of smart home security protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a system architecture schematic diagram of the real-time video behavior recognition and anomaly monitoring system based on edge computing of the present invention;

[0030] Figure 2 This is a flow chart of the real-time video behavior recognition and anomaly monitoring method based on edge computing of the present invention. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] like Figure 1-2 As shown: a real-time video behavior recognition and anomaly monitoring system based on edge computing, including an edge computing node, a video preprocessing module, a background separation and target detection module, an alarm triggering and equipment linkage module, a behavior recognition module and an anomaly monitoring module, the edge computing node is connected to the video preprocessing module and the alarm triggering and equipment linkage module respectively, the video preprocessing module is connected to the background separation and target detection module, the background separation and target detection module is connected to the behavior recognition module, the behavior recognition module is connected to the anomaly monitoring module, and the anomaly monitoring module is connected to the alarm triggering and equipment linkage module;

[0033] Edge computing nodes are integrated into smart home cameras or sensor devices. They are equipped with high-performance processors and dedicated algorithms for local analysis of real-time video streams. Edge computing nodes analyze data collected from terminals directly in local devices or networks close to where the data is generated, without the need to transmit data to cloud data processing centers. This significantly reduces data transmission latency and bandwidth consumption.

[0034] The video preprocessing module performs denoising and format conversion preprocessing operations on video data to ensure the quality and consistency of video data. The video preprocessing module uses image processing technology to improve the clarity and stability of video data, providing a basis for subsequent behavior recognition and abnormality monitoring;

[0035] The background separation and target detection module uses advanced background separation technology to identify moving targets in the video and accurately detect them. The background separation and target detection module uses image processing technology such as background difference method or optical flow method to identify moving targets in the foreground by comparing the difference between the current video frame and the background model, and accurately detects them. Background separation is a common technology that uses a static camera to generate a foreground mask (i.e., a binary image containing pixels belonging to moving objects in the scene) to achieve this;

[0036] The behavior recognition module recognizes the detected target behavior based on machine learning or deep learning algorithms, including specific behavior patterns of people, animals, and vehicles. The behavior recognition module recognizes the target behavior, such as walking, running, and staying, by analyzing the target's motion trajectory, speed, and posture characteristics, using pre-trained models or algorithms. Machine learning is an artificial intelligence technology that can automatically learn patterns and use learning results to predict or make decisions given a data set. It relies on mathematical and statistical algorithms to build models that enable computers to learn autonomously without explicit programming. Deep learning is a machine learning method that uses neural networks composed of multiple levels to simulate and solve complex problems. Its core idea is to gradually improve the model's ability to represent and predict data through hierarchical learning and feature extraction. In deep learning, core technologies include backpropagation algorithms, convolutional neural networks, recurrent neural networks, and generative adversarial networks. Deep learning has a wide range of application scenarios, such as image recognition, speech recognition, natural language processing, and artificial intelligence.

[0037] The anomaly monitoring module monitors abnormal behaviors or events in the video according to preset rules or models. Once an anomaly is detected, an alarm is triggered immediately. The anomaly monitoring module can identify abnormal patterns that do not match the target behavior, such as intrusion, falling, and lost items, and trigger corresponding alarms or notifications. The alarm triggering and device linkage module is linked with other devices in the smart home system to form a complete intelligent security system. The alarm triggering and device linkage module connects and communicates with other devices through the communication protocol or interface of the smart home system to realize the alarm triggering and device linkage functions.

[0038] The present invention also provides a real-time video behavior recognition and anomaly monitoring method based on edge computing, comprising the following steps:

[0039] S1, initialization operation, when the edge computing node starts, it loads the necessary software modules, configures the network connection, checks the hardware status, and prepares the algorithms and models required for video processing. It then checks the connection status to ensure that the video stream can be stably transmitted to the edge computing node, and then sets the alarm trigger conditions, configures the device linkage rules, and defines the recognition criteria for abnormal behavior;

[0040] S2, the edge computing node receives the video stream from the smart home camera in real time through a wireless network or wired connection, monitors the transmission quality of the video stream, ensures that the image is clear, without lag or delay, and temporarily stores the received video stream in the buffer of the edge computing node;

[0041] S3, video preprocessing module preprocesses the video stream, including denoising, format conversion and image enhancement;

[0042] S4, the background separation and target detection module identifies the moving target in the foreground by comparing the difference between the current video frame and the background model, and analyzes the foreground image, identifies the specific moving target, tracks the moving trajectory of the target, and records the position, speed and direction information;

[0043] S5. The behavior recognition module uses machine learning or deep learning algorithms to analyze the target behavior pattern, then extracts key features from the target's motion trajectory and posture changes, matches the extracted features with known behavior patterns, determines the target's current behavior type, and then evaluates the normality or abnormality of the behavior, providing a basis for subsequent abnormality monitoring;

[0044] S6, the abnormal monitoring module detects abnormal behaviors or events in the video according to preset rules or models;

[0045] S7. Once abnormal behavior is detected to trigger an alarm, the user or relevant security personnel will be notified. At the same time, it can be linked with other smart home devices, such as turning on lights, starting camera recording, and sending notifications to mobile phone apps. During this process, relevant information of abnormal events can be recorded, including timestamps, locations, target types, behavior types, and triggered alarms and linkage responses;

[0046] S8. The system will record relevant information of abnormal events. According to the results of abnormal events, the system may update the background model to adapt to environmental changes or improve the accuracy of recognition;

[0047] S9. If no abnormal behavior is detected, continue to monitor the video stream and repeat the process from S1 to S8 to collect new video data and update the internal state.

[0048] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. Real-time video behavior recognition and anomaly monitoring system based on edge computing, characterized by: It includes an edge computing node, a video preprocessing module, a background separation and target detection module, an alarm triggering and equipment linkage module, a behavior recognition module and an abnormality monitoring module. The edge computing node is connected to the video preprocessing module and the alarm triggering and equipment linkage module respectively, the video preprocessing module is connected to the background separation and target detection module, the background separation and target detection module is connected to the behavior recognition module, the behavior recognition module is connected to the abnormality monitoring module, and the abnormality monitoring module is connected to the alarm triggering and equipment linkage module.

2. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 1 is characterized in that: The edge computing node is integrated into a smart home camera or sensor device and is equipped with a high-performance processor and dedicated algorithms for local analysis of real-time video streams.

3. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 1 is characterized in that: The video preprocessing module performs denoising and format conversion preprocessing operations on the video data to ensure the quality and consistency of the video data.

4. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 1 is characterized in that: The background separation and target detection module adopts advanced background separation technology to identify moving targets in the video and accurately detect them. The background separation and target detection module uses image processing technology such as background difference method or optical flow method to identify moving targets in the foreground and accurately detect them by comparing the difference between the current video frame and the background model.

5. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 1 is characterized in that: The behavior recognition module identifies the detected target behavior based on machine learning or deep learning algorithms, including specific behavior patterns of people, animals, and vehicles. The behavior recognition module identifies the target behavior by analyzing the target's motion trajectory, speed, and posture characteristics using a pre-trained model or algorithm.

6. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 5 is characterized in that: Machine learning is an artificial intelligence technique that, given a data set, can automatically learn patterns and use the learning results to make predictions or decisions. It relies on mathematical and statistical algorithms to build models that allow computers to learn autonomously without being explicitly programmed. Deep learning is a machine learning method that uses a neural network composed of multiple levels to simulate and solve complex problems. Its core idea is to gradually improve the model's ability to represent and predict data through hierarchical learning and feature extraction.

7. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 1 is characterized in that: The anomaly monitoring module monitors abnormal behaviors or events in the video according to preset rules or models, and immediately triggers an alarm once an anomaly is detected. The anomaly monitoring module can identify abnormal patterns that do not match the target behavior and trigger corresponding alarms or notifications.

8. The real-time video behavior recognition and anomaly monitoring system based on edge computing according to claim 1 is characterized in that: The alarm triggering and device linkage module is linked with other devices in the smart home system to form a complete intelligent security system. The alarm triggering and device linkage module is connected and communicated with other devices through the communication protocol or interface of the smart home system to realize the alarm triggering and device linkage functions.

9. A real-time video behavior recognition and anomaly monitoring method based on edge computing, characterized in that: The following steps are involved: S1, initialization operation, when the edge computing node is started, the necessary software modules are loaded, the network connection is configured, the hardware status is checked, and the algorithms and models required for video processing are prepared, and then the connection status is checked to ensure that the video stream can be stably transmitted to the edge computing node, and then the alarm triggering conditions are set, the device linkage rules are configured, and the recognition criteria for abnormal behavior are defined; S2, the edge computing node receives the video stream from the smart home camera in real time through a wireless network or a wired connection, monitors the transmission quality of the video stream, ensures that the image is clear, without freezes or delays, and temporarily stores the received video stream in the buffer of the edge computing node; S3, the video preprocessing module preprocesses the video stream, including denoising, format conversion and image enhancement; S4, the background separation and target detection module identifies the moving target in the foreground by comparing the difference between the current video frame and the background model, and analyzes the foreground image, identifies the specific moving target, tracks the moving trajectory of the target, and records the position, speed and direction information; S5. The behavior recognition module uses machine learning or deep learning algorithms to perform pattern analysis on the target behavior, then extracts key features from the target's motion trajectory and posture changes, matches the extracted features with known behavior patterns, determines the target's current behavior type, and then evaluates the normality or abnormality of the behavior, providing a basis for subsequent abnormality monitoring; S6, the abnormal monitoring module detects abnormal behaviors or events in the video according to preset rules or models; S7, once abnormal behavior is detected to trigger an alarm, the user or relevant security personnel will be notified, and other smart home devices can be linked; S8. The system will record relevant information of abnormal events. According to the results of abnormal events, the system may update the background model to adapt to environmental changes or improve the accuracy of recognition; S9: If no abnormal behavior is detected, the video stream is continuously monitored and the process from S1 to S8 is repeated to collect new video data and update the internal state.

Images