CN106375721A - An Intelligent Video Surveillance System Based on Cloud Platform - Google Patents

Abstract

The invention relates to a smart video monitoring system based on a cloud platform, and belongs to the technical field of remote smart monitoring. The system is characterized in that a video monitoring cloud platform is developed on a server cluster established on the basis of a Hadoop cloud computing platform, and video data acquired by front-end equipment are transmitted to a server in a stream way; through adoption of a massive distributed cloud computing technology, real-time tracking, processing, analysis and outputting of video streams are realized; and a service is provided externally, so that a terminal user can perform checking and calling conveniently. In the system, the characteristics of distributed high-concurrence access storage, high availability, rapid response, easiness in management and maintenance and the like are integrated. Video acquisition, video encoding and transcoding, video data storage, a streaming media distributed server and a client are all distributed on a physical host node in an access network. Compared with a conventional video monitoring system, the smart video monitoring system has the advantages that video real-time processing and transcoding efficiency and video data storage capacity are increased greatly; the requirement of people on the video monitoring system in a big data era is met; and a relatively large commercial value is achieved.

Description

An Intelligent Video Surveillance System Based on Cloud Platform

technical field

The invention belongs to the technical field of remote intelligent monitoring, and relates to an intelligent video monitoring system based on a cloud platform.

Background technique

In today's society, with a large population, there are many potential safety hazards. As an important means of security, video surveillance has a large number of cameras distributed in various places. It is unrealistic to rely solely on manual monitoring to generate massive video data. At the same time, video data storage and network transmission bring a lot of pressure. In the modeling process of the video surveillance system, due to system analysis or other reasons, some resources such as cameras and servers may be idle or in short supply. However, various surveillance systems operate independently and lack a platform for effectively integrating resources, resulting in resource allocation. Reasonable, waste a lot of manpower and material resources.

At present, the construction of smart home, smart security and smart city is in full swing. With the increasing scale of video surveillance, there are more and more problems arising. How to use modern communication and network technology to The integration of various security-related video resources in various regions and how to use advanced technologies such as the Internet of Things, big data, and cloud computing to manage video surveillance data in a unified manner have become urgent problems to be solved.

Based on the above background, the present invention designs an intelligent video detection system based on a cloud platform. Aiming at the real-time transcoding processing efficiency and storage capacity of today's video monitoring systems, a cloud computing-based video monitoring system is designed in combination with cloud computing technology. On the one hand, it solves the problem that the limited bandwidth cannot transmit the massive amount of video data in multi-channel monitoring and the system has poor real-time transcoding ability; question.

Contents of the invention

In view of this, the object of the present invention is to provide a cloud-based intelligent video monitoring system, specifically including the following technical solutions:

An intelligent video monitoring system based on a cloud platform, the system includes a front-end acquisition module, a cloud transcoding module, a data storage module, a streaming media module, a video intelligent analysis module, a central scheduling module, a security alarm module and a client;

The front-end acquisition module is used for information acquisition of monitoring video data;

The cloud transcoding module is responsible for transcoding the video information collected by the front end, and providing it to other modules in the system for processing or providing a video format corresponding to the client;

The data storage module is responsible for storing a large amount of video files and key data, providing users with historical video review and related data download services, and meeting the storage requirements of other modules in the entire video surveillance cloud platform;

The streaming module is used to provide RTSP (Real Time Streaming Protocol, real-time streaming protocol) streaming service, and the user obtains the processed real-time video data in real time according to the RTSP address to realize remote monitoring and access, different accesses for users The code stream needs to be provided;

The video intelligent analysis module is used to analyze and process video data in real time, realize the functions of automatic detection, identification, tracking and early warning threat targets, and input the processing results to HDFS storage server or streaming media server in the form of standard streaming media;

The central scheduling module completes the processing (computing and storage) of various tasks by scheduling cloud cluster nodes to realize unified management and monitoring of the entire system platform;

The security alarm module is used to realize that when the network camera detects an abnormal situation, the cloud platform sends alarm information through the wireless network to notify the user terminal of the binding system;

The client provides an operation interface for the user to realize the application of video monitoring, query, historical video playback and remote configuration.

Further, the front-end collection module includes various front-end devices such as network cameras, analog cameras, hard disk video recorders, bayonet devices, video capture cards, embedded IPC high-definition cameras, etc., and collects information through the scheduling of the distributed management module.

Further, the cloud transcoding module adopts the distributed cloud transcoding technology based on Hadoop architecture, puts the processing and calculation of transcoding on the cloud platform, and quickly and efficiently batches the massive surveillance videos collected by the front-end equipment through distributed means. Real-time transcoding can meet the transcoding requirements of various video formats.

Further, the video intelligent analysis module processes the video data from three levels of target detection and tracking, target recognition, and behavior analysis, and triggers the alarm system to notify the SMS and automatically store the abnormal video in a specific location when an abnormal situation occurs. Location.

Further, the central scheduling module schedules each node cloud storage and processing server in the cluster to achieve load balancing through reasonable dynamic node scheduling of the cloud platform, and can only accept subnodes and terminal devices of successfully registered storage or processing servers.

Further, when the client performs real-time video monitoring on the front end, it first interacts with the HDFS distributed video storage server, and then sends the video data stream to the real-time monitoring module in the processing module through the streaming media server.

Further, the cloud platform sends alarm information to notify the user terminal of the binding system through a wireless network, the wireless network includes but not limited to a 4G mobile communication network, and the user terminal includes but not limited to a smart phone and a tablet computer.

The beneficial effects of the present invention are:

1) Remote real-time monitoring and historical video playback

This system controls the storage system and the database (HBase) of the Hadoop cloud platform through the access and control commands generated by JSP to achieve the management and monitoring of video data and cluster device nodes, and can quickly and conveniently combine the server to increase or decrease the number of terminal nodes and so on. A variety of functional servers (video, web, streaming media) are designed in the system server, so that users can check the monitoring video and screen capture of the monitoring terminal through the web page at any time and anywhere in real time at the terminal, and can send information including time, IP or other information commands to query historical videos and perform some processing operations on videos, watch video information collected by a specified terminal, and download historical videos.

2) Intelligent detection and automatic alarm

Users pre-set alarm conditions and rules on the client, and the system can continuously track and detect targets entering the monitoring area according to the conditions set by the client. When the target violates the pre-set rules in the monitoring area, such as intrusion, cross-border, stranded, docked, moved, left over, etc., the on-site situation will be compared and matched with the pre-set conditions, and an alarm message will be sent to the client if they match. And output and store abnormal monitoring video. At the same time, after the cloud monitoring platform sends out the alarm information, it will use various monitoring devices such as network cameras to lock and track the target, and display the status and tracking track of the target. Improve the quality of video surveillance and effectively save human resources.

3) Massive video data storage and processing capabilities

The cloud storage and processing servers of each node in the cluster use the system's reasonable dynamic node scheduling to meet load balancing and coordinate with each other to store and process video stream data efficiently, quickly, and with certain reliability, and only accept registration Subnodes and end devices of successful storage or processing servers. At the same time, the system designs a mode storage interface based on the Adapter adapter, which is convenient for the upper application to call. The intelligent monitoring cloud platform adopts the storage management system based on the ZooKeeper cluster, which can ensure the stability of the storage system. In this way, when a storage node is abnormal or fails, the system can respond in a very short time and add a new node to replace it, so that the video surveillance data can maintain integrity and reliability even when a node is down.

Description of drawings

In order to make the purpose, technical scheme and beneficial effect of the present invention clearer, the present invention provides the following drawings for illustration:

Fig. 1 is a module topological diagram of the present invention;

Fig. 2 is a logical structural diagram of video monitoring software of the present invention;

Fig. 3 is a flow chart of the intelligent video monitoring algorithm of the present invention.

detailed description

The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

Fig. 1 is a module topology diagram of the present invention, as shown in the figure, the intelligent video monitoring system based on cloud computing provided by the present invention adopts the Hadoop cloud computing development framework, which mainly includes the following components: front-end acquisition module, cloud transcoding module, data storage module, streaming media module, video intelligent analysis module, central scheduling module, security alarm module and client.

Specifically:

Front-end acquisition module: including network camera, analog camera, hard disk video recorder, bayonet device, video capture card, embedded IPC high-definition camera and other front-end equipment, through the scheduling of distributed management module to collect information, video information is transmitted in standard RTSP The format is pushed to the cloud transcoding module for transcoding and then input to the cluster storage/processing server for video storage and processing.

Cloud transcoding module: Responsible for transcoding the video information collected by the front-end equipment and providing it to other subsystems in the video surveillance system for calls, such as: storing it in a distributed storage cluster and transferring it to the video surveillance cloud platform for distributed processing. And the transcoding module can also monitor the front-end video signal acquisition equipment in real time. If an abnormality is detected in the front-end video information acquisition equipment, the transcoding system will generate an alarm signal, which can reduce the problems caused by the abnormality of the front-end video information acquisition equipment.

Central scheduling module: As an extremely important control module of this system, it realizes the signaling interaction between the system and the client, completes the processing of various tasks (computing and storage) by scheduling cloud cluster nodes, and realizes unified management and monitoring of the entire system platform. Schedule the front-end equipment according to the requirements, and distribute the JobTracker tasks, process information data, and complete the scheduling work of the entire system platform.

Data storage module: The distributed (HDFS) storage module based on the Hadoop architecture is mainly responsible for the reasonable and reliable storage of massive video data, providing users with historical video review and related data download services, and can meet the needs of the entire monitoring system platform. Storage requirements for other modules.

Video intelligent analysis module: mainly responsible for processing transcoded video information, real-time analysis and processing of video data, automatic detection, identification, tracking and early warning of threat targets, and the processing results are input to HDFS in the form of standard streaming media storage server or streaming server.

Streaming media module: Provide RTSP streaming media service, users can obtain processed real-time video data in real time according to RTSP address to realize remote monitoring and access. Provide code streams for different access requirements of users.

Client: Responsible for the configuration of video information transcoding and storage, management of front-end video information collection, real-time and historical video playback and storage management, cloud monitoring platform control, etc. Real-time monitoring is carried out through webpage access based on the B/S network structure, and direct information interaction with users is carried out.

Security alarm module: When the network camera detects an abnormal situation, the cloud monitoring platform sends an alarm message through the 4G network to notify the user terminal (mobile phone, IPAD, etc.) The role of monitoring.

Fig. 2 is a logical structure diagram of the video surveillance software of the present invention. The intelligent video surveillance software is deployed on the cluster server and includes three parts: a video transmission terminal, an intelligent video analysis module and a monitoring terminal.

The video transmission end is mainly composed of a transcoding module and a storage module in the system. When the video transmission end sends the video data to the intelligent analysis module, the intelligent analysis module will perform processing on the video data such as detection, classification, target tracking, and target behavior recognition according to the request submitted by the client and the video processing setting conditions, and When an abnormal situation occurs, the alarm system is triggered to alarm and the abnormal video is automatically stored in a specific location. The intelligent analysis module is mainly responsible for the real-time processing tasks of the surveillance video data collected by the front-end equipment, and forwards the processing results to the standard streaming media server in real time so that the user can call and view them or send them to the storage module for storage. Several advantages such as rapidity and fast query speed.

Fig. 3 is the flow chart of intelligent video surveillance algorithm of the present invention, and described intelligent video surveillance algorithm can be divided into three levels to the processing of video image, and bottom layer is the detection and tracking of target, middle layer is target recognition, and high level is behavior analysis. The main content of algorithm research is how to extract semantic understanding in line with human cognition from original video data. For example, judging which objects of interest in the scene, historical trajectory, what behaviors to engage in, and the relationship between objects, etc., involves many research fields such as machine vision, pattern recognition, and artificial intelligence. In different application scenarios, for various algorithms, such as SVM support vector machine, GBDT, K-Means clustering, AdaBoost, artificial neural network and other algorithms, the optimal algorithm is selected by testing and comparing parameters after cross-validation Solve the corresponding problems. 00 The monitoring terminal is set on the client side of the entire video monitoring cloud platform. The operator can watch the monitoring video through the web page on the client side and set monitoring conditions, alarm response and operation instructions for the entire system, so as to realize real-time monitoring, remote monitoring and intelligent monitoring .

According to the above content, it can be seen that the functions of the intelligent video surveillance software mainly include:

1) Real-time automatic alarm: the user pre-sets the alarm conditions and rules on the client, and the system will compare and match the on-site conditions with the pre-set conditions when there are unexpected abnormal situations or behaviors (such as: remnants, intrusion detection, etc.) , if they match, an alarm message will be sent to the client and the abnormal monitoring video will be output and stored, which improves the quality of video monitoring and effectively saves human resources.

2) Target detection and tracking: The system can continuously track and detect targets entering the monitoring area according to the conditions set by the client. When a target violates the pre-set rules in the monitoring area, such as intrusion, border crossing, stranded, docked, moved, left over, etc., the status and trajectory of the target will be visually marked, and an alarm will be sent to the cloud monitoring platform. After the cloud monitoring platform locks the target, Send instructions to various monitoring equipment, control the high-speed pan-tilt to drive the tracking camera, realize the locking of the target, and automatically control the pan-tilt and lens parameters according to the movement of the target, always keep the target in the center of the screen, and realize the tracking of the target , and display the trace. The tracking video data and equipment parameters are transmitted to the system cloud platform to realize the detailed display of the target of concern.

3) Automatic video search function: The monitoring system in today's society is huge and extremely complex. Searching for the required monitoring images from a large amount of video data manually requires a huge workload and very low efficiency. The storage method of the system provides an automatic video search function, and the user only needs to remotely submit a request to the server to search for the desired video.

The processing of the video data by the transcoding module includes:

1) For historical videos, the cache server module searches the database for the video and the video format required by the user. If the video exists and the format is exactly what the user requires, the video will be directly returned to the client by the server for its invocation. Otherwise, the video parameter configuration information will be passed to the video transcoding service system and then initiate the transcoding task of the video.

2) For non-historical videos, the user sets transcoding parameters for a video capture terminal (if not, follow the default video transcoding parameter information), and transcodes the video captured by the terminal in real time.

The storage module uses API to input the video data stream files from the local server to the distributed storage system. During uploading, the video data stream files collected by the front-end will be continuously saved in the local video cache directory first, and these data stream files will continue to dynamically increase, using the local directory as a "buffer", and Jobtracker will The specific responsibility of each node in the server connects the HDFS distributed system with the "buffer" in the form of a stream, and then directly uploads the video files on the background server to HDFS in the form of a data stream, achieving the distribution of video data storage purposes. And when the files are successfully uploaded, delete the files in the "buffer" in batches.

The streaming media module adopts the framework of the DSS streaming media server to design and develop, and the video data is packaged by RTP/RTCP to realize the mutual transmission between the modules through the network, and the client uses the RTSP/SDP protocol to control the streaming media The operation of the server. After transcoding, the video data collected by the front-end will be sent to the real-time monitoring and historical video query and call frameworks through the streaming media server for specific implementation and processing, and can be returned to the client according to different requests provided by the web client. or the historical video data after being processed by the media file framework as a response, realizing the call playback of historical surveillance video. When the user sends a request to watch a video in real time on the client side, the streaming media server will implement the real-time playback of monitoring video and some basic operations on the client side through the standard real-time streaming protocol.

The functions of the client module are:

1) User information management, the client software will grant users different permissions and access permissions based on the information on the user table, manage user login, add and delete users with different permissions.

2) Camera management and maintenance, using a tree structure management method to manage the front-end camera equipment, manage the operation status of the front-end video surveillance equipment and nodes according to the current user's authority, and update the list in time.

3) Real-time monitoring, scheduling and dispatching tasks through JobTracker, and monitoring the execution of tasks to obtain real-time video streams to achieve real-time monitoring.

4) Historical video search and playback, by sending a request to the streaming media server, the address is obtained to download the historical video on the storage server for playback processing.

5) Event management, target detection, object tracking, intelligent recognition, etc. in video surveillance.

The intelligent monitoring cloud platform adopts a storage management system based on ZooKeeper, which can ensure the stability of the storage system. When a storage node is abnormal or faulty, the system can respond in a very short time and add a new node to replace it, so that the video surveillance data can maintain integrity and reliability even when a node is down, and The efficiency and performance of video storage in the MapReduce system is not compromised by backup and redundant video files.

The intelligent monitoring cloud platform is developed on the B/S using the JSP web page design technology and the web management function that comes with the Hadoop framework. The system controls the storage system and the database (HBase) of the Hadoop cloud platform through the access and control commands generated by JSP to achieve the management and monitoring of video data and cluster device nodes, and can quickly and conveniently combine the server to increase or decrease the number of terminal nodes, etc. operate. A variety of functional servers (video, web, streaming media) are designed in the system server, so that users can check the monitoring video and screen capture of the monitoring terminal through the web page at any time and anywhere in real time at the terminal, and can send information including time, IP or other information commands to query historical video and perform some processing operations on the video, watch video information collected by a specified terminal, and download historical video.

Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that it can be described in terms of form and Various changes may be made in the details without departing from the scope of the invention defined by the claims.

Claims (7)

1. a kind of intelligent video monitoring system based on cloud platform it is characterised in that: this system includes front end acquisition module, high in the clouds Transcoding module, data memory module, streaming media module, video intelligent analysis module, central dispatching module, security alarm module and Client；

Described front end acquisition module is used for carrying out information gathering to monitor video data；

Described high in the clouds transcoding module is responsible for for the video information of front-end collection carrying out transcoding, and is supplied at other modules in system Manage or the video format corresponding with user side is provided；

Described data memory module is responsible for storing video file and the critical data of magnanimity, provides the user returning of history video See, related data download service, meet the storage demand of other modules in whole video monitoring cloud platform；

Described streaming media module is used for providing rtsp (real time streaming protocol, real time streaming transport protocol) to flow Media services, user according to rtsp address, the real time video data that acquisition in real time was processed realizing remote monitoring and access, There is provided code stream for the different requirements for access of user；

Described video intelligent analysis module is used for video data is analyzed and processed in real time, realizes automatic detection, identification, follows the tracks of And early warning threatens the function of target, and result is input to hdfs storage server or stream in the form of standard Streaming Media Media server；

Described central dispatching module converges, by scheduling, the process that group node completes each task, realizes unified management and monitors Whole system platform；

Described security alarm module is used for realizing when IP Camera detects unusual condition, and cloud platform is sent out by wireless network Warning information is sent to notify the user terminal of binding system；

Described client provides the user operation interface, realizes the monitoring to video, inquiry, history video playback and Remote configuration Application.

2. a kind of intelligent video monitoring system based on cloud platform according to claim 1 it is characterised in that: described front end Acquisition module includes web camera, analog video camera, DVR, tollgate devices, video frequency collection card, embedded ipc height Multiple headend equipments such as clear video camera, carry out information gathering by the scheduling of distributed management module.

3. a kind of intelligent video monitoring system based on cloud platform according to claim 1 it is characterised in that: described high in the clouds Transcoding module, using the distributed high in the clouds transcoding technology based on hadoop framework, the process of transcoding and calculating is placed on cloud platform On, quickly and efficiently batch real-time transcoding is carried out to the magnanimity monitor video of headend equipment collection by distributed means, can Meet the transcoding demand of various video form.

4. a kind of intelligent video monitoring system based on cloud platform according to claim 1 it is characterised in that: described video Intelligent analysis module, is processed from target detection tracking, target recognition, three levels of behavior analysiss to video data, and And triggering warning system carries out note notice and automatically anomalous video is stored in ad-hoc location when abnormal conditions.

5. a kind of intelligent video monitoring system based on cloud platform according to claim 1 it is characterised in that: described center Scheduler module dispatches the rational dynamic node scheduling passing through cloud platform in cluster between each node cloud storage and processing server To realize load balancing, and child node and the terminal unit of the storage or processing server succeeding in registration can only be accepted.

6. a kind of intelligent video monitoring system based on cloud platform according to claim 1 it is characterised in that: described client When end carries out real-time video monitoring to front end, interact with hdfs distributed video storage server first, then taken by Streaming Media Business device by video data stream, to realize by the real-time monitoring module in processing module.

7. a kind of intelligent video monitoring system based on cloud platform according to claim 1 it is characterised in that: cloud platform lead to Cross wireless network and send the user terminal that warning information notifies binding system, it is mobile logical that described wireless network includes but is not limited to 4g Communication network, described user terminal includes but is not limited to smart mobile phone and panel computer.