Intelligent Video Identification and Analysis System
Industry Problems
The development of video monitor industry has experienced 3 stages, from the original stage of analog video monitor, to the stage of partial digital monitor, to the nowadays stage of network digital video monitor. However, there are shortages in no matter what stages:
•     If the monitor totally relies on manuallabor, the efficiency of the whole video monitor system cannot be ensured.
•     The manual labor monitor time have influence, each monitor point is under the observation of monitor people for only a short time, and the probability of missing report is very high.
•     The data analysis is very difficult, the only solution is replaying the video and using manual labor to retrieval, and the reliability is very low.
•     The responding time is long, and lack of real time efficiency, cannot warn in time and precisely.
Solution Overview

CertusNet intelligent video analysis platform is a set of advanced video content analysis system, and it is used for video tracking and behavior analysis of the object in the video frames. The system is based on high level video diagnose algorithm which is independent researched and developed. The system can parallel calculate and process video content in high efficiency, and realize multiple kinds of survey and tracking for the object. The system can deal with intervene in all kinds of scenes and environment, and has very low error reporting rate. 
 

Feature
•     Video Source Support
DVR/DVS digital video
IP video cameral
Video capture card
Video monitor platform
Video file
•     Video Analysis Function
Every equipment support 16 lines video insertion
Each line of video can set at most 10 warning types
Each line of video can detect 20 objects at same time
The least detected object can be 5*5 pixels
The lag of video is less than 2s
•     Video Analysis Function
Detection range can define the range of full screen, rectangle, andirregular.
Classify the objects as people and cars etc.
Filter disturb caused by little animals, shaking trees, shade of cloud, light change of scenes, camera tilting.
•     Video Analysis Type
Invasion detection
Railing detection
Detection of entrance and quit of objects
Wandering detection
Remnant detection
Object behavior detection (emerging, disappearing, and moving)
People crowd density detection
Running detection
Reversing detection
Flow rate of people and vehicle detection
Road condition detection
Vehicle violation parking detection
Scene change detection

 

Key Technologies
•     Object extraction algorithm technique
•     Space model of moving objects technique
•     Kalman filter algorithm
•     Retrieval and tracking technique of moving object based on active vision
•     H.264 coding technique
Value
Compare with traditional video monitor, intelligent video monitor change the passive monitor way to the initiative way. It can warn and provide useful information in the quickest and fastest way. It’s main advantages shows as follows:
•     Intelligent video monitor system can continuously analyze monitor pictures 7*24 hours according to rules set by users.
•     The lag of intelligent video monitor is very low. The intelligent video monitor can alarm at the first time of unusual condition, alarm the monitor people to pay attention. Intelligent video monitor can inform other systems to respond united at real time, and greatly lessen the responding time of the whole security defense system.    
•     Intelligent video analysis system equipped and intelligent algorithms, can effectively remedy the insufficient of manual labor monitor, and reduce the error reporting rate and missing reporting rate of the whole video monitor system.
•     Intelligent video monitor system has many utility other than security object. For example, people counting function based on intelligent video analysis. This function can be used in business places to count the customer number of each day, and the date collected has certain business reference value.