a monitoring method

Abstract

The present invention provides a monitoring method, comprising: acquiring a shot video frame sequence; extracting the foreground image of the current video according to the video frame sequence based on a pre-trained neural network model; removing the foreground image according to a preset shadow removal method The shadow in the foreground image; determine whether there is an abnormal hole in the foreground image, and if so, fill the abnormal hole according to the preset abnormal hole filling method; obtain the circumscribed rectangle of the foreground image, and highlight the circumscribed rectangle. The invention can quickly and accurately obtain the exact shape of the non-background image in the video picture, thereby realizing functions such as comparison of suspects, display of key pictures, and display of the center of the picture. The invention has high intelligence, high accuracy of acquired graphics and broad application prospects.

Description

technical field [0001] The invention relates to the monitoring field, in particular to a monitoring method. Background technique [0002] The intelligent monitoring system adopts image processing, pattern recognition and computer vision technology. By adding an intelligent video analysis module to the monitoring system, the powerful data processing capability of the computer is used to filter out useless or disturbing information in the video screen, automatically identify different objects, analyze and extract The key and useful information in the video source can quickly and accurately locate the accident scene, judge the abnormal situation in the monitoring screen, and issue an alarm or trigger other actions in the fastest and best way, so as to effectively carry out pre-warning, processing during the event, and timely after the event Fully automatic, all-weather, real-time monitoring intelligent system for evidence collection. [0003] In the prior art, intelligent moni...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, intelligent monitoring has been widely used, but the relevant technologies for extracting images from flowing video frames and automatically processing the extracted images are not yet mature, so it is difficult to achieve complete automation, and it still needs to rely on Identification with the naked eye

In view of the immature technology of automatic image processing, the false alarm rate of automatic alarm based on image processing technology is high, and it is difficult to obtain the accurate outline and precise position of the suspected target

Images