Product Appearance
Product Overview
The spatio-temporal analyzer professional edition, based on the video structuralization algorithm of deep learning, comes as an upgraded version of the standard edition, whether in function or hardware. GPU processor is added to raise the efficiency and accuracy in handling the video, realize an extraction of the moving objects in the video followed with a structuralized analysis of the relevant person, vehicle and things, and rapidly obtain the valuable information in the video through such means as reverse image search and structuralized feature search.
Function Overview
Information Management
The system supports description of cases, and description and management of video point location information, including the corresponding coordinates and collection time.
Structuralized Video Processing
The system supports a structuralized processing of such information as human face, human body and vehicle in previous video documents.
Structuralized Video Retrieval
The system supports a quick retrieval of the structuralized features of the relevant person, vehicle and things (such as human face, gender, hat, texture, vehicle type, color and model, sunvisor and annual inspection standards).
Reverse Image Search
The system allows the user to select the to-be-checked object following the automatic image recognition, and search the to-be-checked video for the similar object through reserve image search, with an automatic sequence by similarity.
Datamation Display
As for information discovered in the course of solving a case, including various view clues and point-in-time, camera and geographical location, the system can offer a graphical presentation of the connections among these clues to directly display them on the interface based on operations like sifting, analysis and combination.
Trace Study & Judgment
Marks of the location, angle, direction and other information about the camera on a map in offline mode, coupled with the time when the suspicious object appeared in each monitory point, are useful in exhibiting clearly the spatio and temporal trace of the object concerned before and after the incidence of a case.
Associated Comparison
The system can play up to four original videos at the same time and make a comparison, and control the display of each video through fast-forwarding, rewinding, forwarding frame by frame, rewinding frame by frame, and so on.
Compressed Display
Setting of the area of interest is allowed in the spatio-temporal analyzer professional edition. The video can be played at an up to 16X speed when there is no moving object in such area; and it can be played at a normal or slow speed when the moving object passes such area.
Image Processing
The system can deliver clear, sharp images through image processing, aside from basic image enhancement, geometric correction, image operation, elimination of noise interference, color interference correction, defuzzification processing, single frame super-resolution amplification, environmental interference correction, feature viewing and analysis, multi-frame sequential processing, 2D/3D figure reconstruction, among other functions.
Case Video Non-linear Editing
The system supports a mark and automatic tracking of the object appearing in the video of a case, and corresponding video trimming and stitching.
Case Reporting
Based on the investigation process, the system is able to automatically generate a case report with the format as specified, which includes basic information about the case, video resources used in the case, monitory point information, case clues and conclusion, among other content.
Product Features
High processing speed: The hardware design of double-GPU boasts a video decoding and processing efficiency three times higher than that of CPU, and is more suitable for processing the current 1080P high-resolution video;
High accuracy: The internationally leading deep learning algorithm for video structuralization is adopted, which greatly improves the accuracy of structuralized feature search and reverse image search. Besides, the search accuracy can also be increased with more pictures;
High compatibility: The system is compatible with various sorts of non-standard videos, and can upgrade the decoding library in time and search for moving objects in videos of a non-ideal scenario.
