Intelligent upgrade shortcut video analysis module features and applications

The solidification and modularization of intelligent video analysis algorithms are the current trends in the application of intelligent video analysis technology. It perfectly combines intelligent video analysis technology with an efficient digital signal processor (DSP), and can embed intelligent video technology into cameras, DVRs, and any network video surveillance products, making them immediately upgraded to intelligent video surveillance solutions, greatly increasing the added value and economic benefits of the products. Its high-performance processors, embedded platforms, and powerful intelligent video analysis algorithms are a shortcut to the intelligent upgrade of DVR/DVS/IPCAM products.

The following is a brief introduction to the features and applications of the intelligent video analysis module with unified functions, productization and modularization.

The intelligent video analysis module can be seamlessly connected with DVR/DVS/IPCAM, complementing each other to form a complete network platform intelligent video surveillance system. The intelligent video analysis platform can also capture information that needs attention from the back-end centralized video source. This method is more suitable for projects such as "Safe City" and other large-scale centralized video surveillance platforms for intelligent transformation.

Examples of Intelligent Video Analysis Technology Applications

1. Dynamic face capture device

Dynamic face occlusion detection uses intelligent video analysis technology, namely the face occlusion detection algorithm, to analyze and detect people's faces and automatically determine their facial features. It can also automatically identify and alarm suspicious behaviors such as wearing sunglasses, lowering the brim of a hat, wearing a mask, or any other form of intentional obstruction of facial features.

It can intelligently separate and capture faces from the back-end centralized video source. By introducing the video source at the back-end of the centralized video surveillance platform, face capture is achieved on the back-end intelligent video analysis platform. Its main performance is as follows:

Automatically detect and capture faces in video scenes. Based on the DSP platform, it supports 1-way or 2-way video input, and can capture faces in video scenes and archive photos. It can provide standard SDK interface protocols, and can perform two-way communication through RJ45, RS232485 and other interfaces, making it easy to integrate with other devices.

Features:

Embedded processing, automatically detecting faces in video scenes, capturing partial photos of faces, and transmitting them to the backend;

Can be directly connected to ordinary analog cameras or digital video streams, and can automatically capture all faces entering and leaving the monitored area 24 hours a day;

The product can be upgraded directly through the network port/U5B interface, which is convenient and fast to maintain; it provides various hardware interfaces such as RS232/RS485, Gang45, U58, SD, etc.

Supports facial photo comparison with the backend and can be used with the software;

Intelligent analysis results: saved to the server through middleware or provided to third parties for use;

Intelligent video recording: The video file can display the results of intelligent analysis through the FIRS player (face size requirement - not less than 100×80 pixels).

Applications:

At the entrances and exits of various places where people gather, such as subways, railway stations, long-distance bus stations, docks, airports, government departments, and halls, automatic face detection and photo capture and archiving are performed on people passing through for later face retrieval and blacklist/whitelist comparison.

It should be said that the functions and main parameters of today's intelligent video analysis modules applied to face capture devices are similar, and the above explanation only takes a certain type of face capture device as an example.

2. Intrusion detection warning device

A device that applies video behavior analysis technology, which can perform large-scale pre-alarm detection, has a high detection rate and a low false alarm rate. The so-called video behavior analysis technology is to detect, classify and track the trajectory of relevant moving targets (people or objects) in the video scene, and according to the established analysis (trigger) rules, the system automatically analyzes and judges the behavior information of the moving targets, and outputs the information to the relevant system platform. The terminal is based on the DSP platform and can have 1 or 2 video inputs. Through intelligent video analysis, it can realize early warning of perimeter intrusion behaviors in the video scene, provide local IO output, and realize local and remote alarm. The terminal provides a standard SDK interface protocol, which can perform two-way communication through RJ45, RS232/485 and other interfaces, and is used for integrated applications with other devices.

Features:

Embedded processing terminal, automatically detects suspicious perimeter intrusion in the video scene, automatically issues local and remote alarms, and preserves evidence; can be directly connected to analog cameras or IPCAMs, automatically performs 24-hour uninterrupted detection. The product can also be upgraded directly through the network port/USB interface; can provide various hardware interfaces such as RS232/485, RJ45, USB, SD, etc.

Supports intelligent intrusion detection such as crossing the line, entering or leaving the restricted area, going in the opposite direction, wandering, etc.; the results of intelligent analysis can be seen in real time, and the detection targets can be marked and tracked.

Applications:

It is installed in various places where perimeter intrusion needs to be prevented, such as residential areas, villas, factories, mines, warehouses, cargo yards, campuses, enterprises, and military bases. It can provide 24-hour uninterrupted detection and early warning for various perimeter intrusion behaviors, such as crossing the line, restricted areas, and reverse driving.

3. Crowd concentration detector

Crowd density detection is also one of the intelligent video analysis technologies. It collects video information through cameras and uses intelligent video analysis technology to analyze the living moving targets in the picture, and then calculates the number and density of targets in the set area. When the number value exceeds the preset threshold, an alarm message is automatically generated to prompt the monitoring personnel to take targeted measures. It can be applied to scenes such as illegal crowd gatherings and airport verification halls.

Features:

Embedded processing, automatically detect suspicious crowd gathering behavior in video scenes, automatically issue local and remote alarms, and preserve evidence; can directly connect to ordinary analog cameras or digital video streams, and automatically perform 24-hour non-stop detection and warning. The product can also be upgraded directly through the network port/USB interface; it can also provide various rich hardware interfaces such as RS232/485, RJ45, USB, SD, etc. The system is based on the DSP platform, automatically analyzes the crowd gathering behavior in the video scene, and issues graded warnings, provides local IO output, and can realize local and remote alarms. It can provide standard SDK interface protocols, and can be integrated with other devices through interfaces such as RJ45, RS232/485, etc.

Applications:

It is suitable for municipal squares, airport and wharf passages, gates of enterprises and institutions, and around military restricted areas, etc., for 24-hour uninterrupted detection and early warning of suspicious crowd gatherings.

Other applications

Intelligent video analysis technology has many other applications, such as video passenger flow statistics, multi-light source face recognition, partial high-definition face capture, dynamic license plate capture, moving target detection, object status detection, camera health detection, etc.