Pedestrian target false detection removal preprocessing method

Abstract

The invention discloses a pedestrian target false detection removal preprocessing method, which belongs to the field of pedestrian detection preprocessing and is used for eliminating common false detection phenomena of pedestrian detection results in an image. According to the technical key points, the method comprises the steps of S1, performing normalized width processing; S2, performing histogram specification equalization processing; and S3, performing sharpening processing. The method has the effects that redundant pedestrian detection boxes appearing in a pedestrian detection process areremoved; the effectiveness of subsequent pedestrian tracking algorithm application is improved; and additional judgment and calculation consumption of a pedestrian detection system due to false detection of targets is avoided.

Description

technical field [0001] The invention belongs to the field of pedestrian detection preprocessing, in particular to a method for reducing the number of false pedestrian targets detected in an image. Background technique [0002] With the rapid development of robot applications, the demand for intelligent robots is also increasing. In recent years, pedestrian detection technology has become one of the hot research directions, and it is also an important part of the field of computer vision. Commonly used pedestrian detection methods are to simulate the human eye, realize the machine to collect images of the surrounding environment or things, and then analyze and identify the collected image content to complete the machine's perception of the three-dimensional world. [0003] At present, the detection ideas of mainstream pedestrian detectors are mainly divided into two categories: machine learning detection methods and deep learning detection methods. The common point of these...

AI Technical Summary

Problems solved by technology

The Chinese invention patent application with the application publication number CN106650667A discloses a pedestrian detection method and system based on support vector machine, which mentions that the support vector machine cannot control the distribution of feature weights well and reduce the accuracy of pedestrian target detection, and proposes a simulation The genetic algorithm of the annealing acceptance criterion comprehensively considers the relationship between the features in the image, and formulates weights for the features of all dimensions, thereby improving the accuracy of pedestrian detection; according to the Chinese invention patent application with the application publication number CN105913003A, a multi-feature multi- Model pedestrian detection method, which analyzes the advantages and disadvantages of various pedestrian detection methods, proposes a multi-feature multi-model pedestrian detection method, integrates the content of various pedestrian detectors, and achieves compatibility with high detection rate and low error at the same time purpose of reporting

Images