Crowd scale statistical method based on color and HAAR feature fusion

Abstract

The invention discloses a crowd scale statistical method based on color and HAAR feature fusion. The method includes the following steps of data preprocessing, feature extraction, objective model establishment conducted through Adaboost, pedestrian detection, and people counting. According to the method, the color feature and the HAAR feature are fused, color, as prior knowledge, serves as the weight of the HAAR feature for objective judgment, and detection and recognition of individuals in a crowd are conducted through an Adaboost training classifier. An HAAR model is trained firstly, the classifier is established through weighing and based on the color attribute of the human face complexion and is used for human face detection, K-NN classification is conducted on human face sub-windows, and a classification result is the number of pedestrians. According to the method, the influences caused by factors such as weak light, objective concentration and too small objectives can be avoided.

Description

technical field [0001] The present invention relates to a scale statistics method, in particular to a crowd scale statistics method based on the fusion of color and HAAR features. Background technique [0002] Video crowd size statistics use video image analysis technology for crowd flow statistics. Analyzing and counting crowd size (that is, the number of people) from video streams is a very complex and challenging computer vision and computational intelligence problem. Crowd size statistics can provide effective data support for the management and decision-making of large shopping malls, shopping centers, chain stores, squares, streets, airports, stations, museums, exhibition halls and other public places. While mastering the dynamic information of the monitoring area, the user can timely obtain the accurate number of people and crowd flow data in the target area, which is conducive to improving the management level and efficiency. The realization of video crowd size sta...

AI Technical Summary

Problems solved by technology

However, through local or global feature fusion, it is difficult to distinguish similar features, which is easy to cause false detection

[0006] Beijing Vimicro Electronics Co., Ltd. Huang Ying's patent "People Counting Method and System Based on Video Surveillance" was applied for and approved by the State Intellectual Property Office of China on January 7, 2009, and was published on January 8, 2009. Publication No. For: CN101477641, this patent mainly uses tracking to count the flow of people. When there are too many people, the following problems will occur: First, problems such as overlap and occlusion between monomers are prone to occur in the process of multi-target tracking, and it is impossible to ensure the tracking required. Acquisition of individual features, so that it is impossible to detect and track; third, multi-target tracking requires a lot of calculations, and the time spent is difficult to estimate

[0008] This article uses Haar features to count the number of people in the image, but the experimental samples used in this article are different from the images obtained by video surveillance. Simply using Haar features to count the number of people in the surveillance will cause missed detection and false detection. At the same time, this article does not address Duplicate counting strategy for video occurrences

Images