Visitor flowrate statistic method

Abstract

The invention discloses a visitor flowrate statistic method, and belongs to the technical field of video monitoring and image processing and analysis. The method comprises the following steps of: calculating gradient information of an input video image; carrying out background updating and gradient foreground extraction on input data which takes the gradient information as a background difference; carrying out gradient weighted Hough transformation on an obtained gradient foreground picture; determining each head in the current image; tracing the detected head and detecting the head again by adoption of a head general classifier so as to form a movement confidence track of a head target; self-adaptively setting a target filtration and fusion threshold value according to the size and movement direction of the head in the movement process; and finally carrying out visitor flowrate statistic according to the evaluation and direction of the movement confidence track evaluation of the head target. The visitor flowrate statistic method provided by the invention is stable and correct in counting and is capable of achieving an ideal statistic effect under various conditions.

Description

Technical field [0001] The invention belongs to the technical field of video monitoring and image processing and analysis, and is specifically a method for counting the flow of people. Background technique [0002] With the continuous progress of society, the application range of video surveillance systems is becoming wider and wider. Surveillance cameras are usually installed at the entrances and exits of shops, shopping malls, subway stations, train stations, airplane stations, stadiums and other places, so that security personnel and management personnel can monitor the above places; at the same time, the flow of people in the above places is important for later decision-making It has a very important role, such as the positioning of hot-selling products in shops, the location of goods, and so on. [0003] At present, there are three main types of traffic statistics methods based on video analysis: one is based on feature point tracking. The disadvantage of this method is that ...

AI Technical Summary

Problems solved by technology

[0003] At present, there are three main types of traffic statistics methods based on video analysis: one is the method based on feature point tracking, the disadvantage of this method is that the feature point itself is difficult to extract and track stably, and the counting accuracy is poor; , the disadvantage of this method is that when the human body is occluded or similar objects interfere, the accuracy of human body segmentation is difficult to guarantee, and the statistical accuracy is unstable; the third is the method based on head or head and shoulder detection and tracking. When different types of targets are detected, the detection rate will decrease, and the derivative measurement accuracy will decrease

Images