A Method for Recognizing Human Objects in Self-directed Images

Abstract

The invention provides a method for identifying human body targets in self-facing images. The video stream image of the monitoring area is obtained as an input image; the monitoring area is planned into a plurality of sub-areas with equal areas; Sampling is performed at a set interval in the sub-region to obtain a number of standard human body model images and perform density difference operations on them and the background, obtain and store the normal vector group representing the outline of the standard human body model as standard reference data; the input image and the background Perform density difference operation to obtain the normal vector group representing the target contour in the input image, determine the response points in the normal vector group of the target contour in the input image in combination with the standard reference data, and judge whether the target contour is Pedestrian silhouette. The method shown in the present invention significantly improves the recognition efficiency of pedestrian targets.

Description

technical field [0001] The invention relates to the technical field of video detection, in particular to a method for identifying a human body target from an opposing image. Background technique [0002] With the development of electronic science and technology, human body detection technology based on image or video has gradually replaced traditional infrared, radar and other human body detection methods. Moreover, human body detection technology based on images or videos has higher reliability. In addition, due to its low-cost characteristics, it has attracted more and more attention. [0003] Fourier transform is a texture analysis method for two-dimensional grayscale images. Fourier transform is usually used in image processing to analyze image texture. The image is divided into square gridded areas, and the two-dimensional grayscale image is detected and analyzed in terms of lattice and defects based on the phase of the square gridded areas. [0004] Since the Fouri...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to disclose a method for identifying human targets from facing images, which is used to realize the density difference of various features of human target outlines in the background image, thereby efficiently extracting pedestrian targets from the background image and overcoming image The input device is unable to effectively recognize the defects of pedestrian targets in complex environments caused by external interference or defects of the lens itself, so as to improve the recognition rate of pedestrian targets

Images