Video Quality Monitoring System

In many companies surveillance security cameras require constant human control of generated video flows or image sequences to monitor situations when a controlled image is not of acceptable quality in definition of the term quality within the requirements of standards. At the same time, the participation of personnel in routine monitoring is expensive and depends on concentration and attention.

Development of the customized processing algorithms, using elements of mathematical statistics and signal processing theory; - Implementation of automatic methods of video quality control to release personnel from routine activities. Since different types of flaws are possible the best solution includes two methods: - Video quality score with machine learning detection Measuring the relative amount of information in each frame - Computer vision (CV) - Detection of scene change - Detection of moving objects on the scene - Classification of moving objects - Detection of face(s) on image - Detection of textual information, OCR detections.

Lean Monitoring - Iterative processing of cameras to detect flaws with a reduced amount of checks System Architecture - Video Quality Monitoring System is integrated with existing Video Control Systems - DATA FLOW IN VQET - SCORING ALGORYTHM - Integrated notification and flaws resolution control service - AI-Module detects “Suspected flaw” and starts separate threads of analysis using computer vision-based technics. - CV-Modules can be added on the later phases, that will use different available methods to classify flaws and give even better results or take actions