Method and system for detecting pedestrian flow by adopting deformable two-dimensional curves

Abstract

The invention provides a method and system for detecting pedestrian flow by adopting deformable two-dimensional curves. The method comprises the following steps: adopting a camera, a video line, a video acquisition card and a computer (or digital signal processing (DSP)-based video developing board) to set up a system hardware platform and manufacturing a software interface; firstly placing the camera right above a door (channel) to vertically shoot coming-in and going-out actions of the pedestrians downwards, acquiring the color image of each frame and extracting pedestrian heads according to red, green and blue (R, G, B) values of the acquired current frame image; detecting the edge contour of each pedestrian head on the current frame image by adopting the image segmentation technology and then using the deformable two-dimensional curves to fit the edge contour of each pedestrian head on the current frame image by adopting the curve fitting technology to compute the parameter and central coordinate of each two-dimensional curve; comparing the parameter and distance between centers of each two-dimensional curve on the previous frame image with the parameter and distance between centers of each two-dimensional curve on the current frame image to find out the correspondence; and finally counting the numbers of pedestrians who come in and go out according to the positional relationship between the corresponding two-dimensional curves on the previous frame image and the current frame image and a counting line.

Description

technical field [0001] The invention relates to a method and system for detecting people flow using a deformable two-dimensional curve, which is used in the application fields of shopping malls, roads, stations and other public places management, non-rigid body target recognition and tracking, video motion analysis, etc., in video image processing, Computer vision, pattern recognition, and artificial intelligence all have practical application value. It belongs to the field of computer vision and video image processing. Background technique [0002] At present, a series of research work has been carried out at home and abroad using human head models for tracking and monitoring. These techniques can be divided into two categories: 2D model-based techniques and 3D model-based techniques. Two-dimensional methods include: edge-based, facial feature-based, facial feature-based, and similarity-based techniques. 3D models include: cylinders, ellipsoids, and polygonal meshes. ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem that the rigid body model cannot accurately represent the shape of the human head in the detection of the flow of people, and propose a method of detecting the flow of people using a deformable two-dimensional curve to detect the human head in the image. Flow detection, reducing the cost of people flow detection, suitable for detection and counting of various moving objects

Images