Dynamic recursion modeling and fusion tracking method for passage blockage judgment

Abstract

The invention belongs to the technical field of computer vision, and in particular relates to a dynamic recursion modeling and fusion tracking method for passage blockage judgment. The method includesthe steps of adopting a dynamic recursion background modeling algorithm for a passage background model to detect the background and the foreground information of a passage, and separating dynamic andstatic objects including pedestrians and stacked objects in the passage, timely and accurately tracking the objects by using a feature fusion tracking algorithm in combination with object colors andposition features, extracting the position characteristic information of the objects, determining whether the objects are in a moving state or a static state, and adopting a background local conditionupdating strategy for different states, determining whether the object is a blocking object or not according to the retention time and the width ratio of the tracked objects in the passage. Accordingto the method, object blockage in a passage can be effectively detected, and the safety guarantee capability of units, where fire disaster is likely to occur frequently can be improved. The potentialsafety hazard problem that an evacuation passage and an emergency exit in a fire disaster high-risk unit are occupied is solved.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to a dynamic recursive modeling and fusion tracking method for channel blockage discrimination. Background technique [0002] With the progress of society, people continue to pursue more comfortable and ample housing and work space, resulting in more and more evacuation passages and safety exits being occupied. Once a fire occurs, it will seriously affect the evacuation and fire rescue, causing a large number of casualties and Property loss has aroused more and more people's attention. However, channel blockage detection cannot be realized by manual inspection alone. Although the relevant departments currently deploy cameras to monitor the corresponding key channel intersections, it is still necessary to manually judge whether the channel is blocked based on the surveillance video, which has not fundamentally solved channel blockage detection. question. [0003...

AI Technical Summary

Problems solved by technology

A fire exit obstacle detection method based on image matching. It only needs to match the background image with the real-time monitoring image in the designated area to judge whether there is an obstacle blockage in the designated area. It is simple to implement, but the background image cannot be updated from time to time. , can not be well adapted to the changes in the complex environment of fire exits; a safety detection method for fire exits based on adaptive background learning, which uses the foreground target detection of HOG features and the adaptive background update strategy, can well adapt to the complex environment of the passage Changes to improve the accuracy and robustness of object detection, but do not consider short-term objects appearing in the channel, which are not blocked; there is also an obstacle detection method technology based on multi-sensor information fusion, which combines millimeter wave Radar and computer vision technology have greatly improved the reliability of channel obstacle detection, but the implementation cost of this method is high, which is not conducive to large-scale promotion

Images