Far-infrared pedestrian training method of gradient magnitude distribution gradient orientation histogram

Abstract

The invention discloses a far infrared pedestrian training method of a gradient magnitude distribution gradient orientation histogram. HOG (Histogram of Oriented Gradient) characteristics widely applied to image classification at present are taken into consideration so that gradient amplitude distribution of images is described yet.. Therefore, HOG characteristics are improved, description of gradient amplitude distribution is designed, the characterization capability of the HOG characteristics is enhanced, and a new gradient orientation histogram characteristic based on gradient amplitude distribution is obtained and is called coded HOG (Coded HOG, CHOG). Training samples are collected for vehicle-mounted far infrared pedestrian classification, and CHOG features of the training samples are extracted. A CHOG feature-based linear support vector machine classifier model is acquired. The method comprises a feature improvement module, a feature extraction module and a CHOG feature-based linear support vector machine training module, wherein the feature improvement module is used for improving the current HOG feature by additionally describing gradient amplitude distribution. The feature extraction module is used for extracting CHOG features of a training sample;. Both the detection and classification accuracy and the classification speed can be considered, and the obtained designedclassifier is the core of the night aided driving pedestrian reminding system.

Description

technical field [0001] The invention belongs to the fields of computer vision and pattern recognition, image processing and computer vision, and in particular relates to a far-infrared pedestrian training method based on a gradient amplitude distribution gradient orientation histogram. Background technique [0002] Far-infrared imaging is based on temperature differences, and is especially suitable for detecting living objects in front, such as pedestrians. At present, the assisted driving system based on the vehicle-mounted far-infrared camera for pedestrian detection is becoming more and more popular. The core component of this type of system is usually a vehicle-mounted far-infrared pedestrian classifier. The performance of this core component directly determines the performance of the overall assisted driving system. However, the design of a robust and real-time far-infrared pedestrian classifier is a challenging task due to the large differences in the clothing of road...

AI Technical Summary

Problems solved by technology

[0011] The purpose of the embodiments of the present invention is to provide a far-infrared pedestrian training method based on the gradient amplitude distribution gradient orientation histogram, aiming to solve the unsatisfactory recognition accuracy and real-time performance of the existing far-infrared camera-based vehicle-mounted pedestrian classification method. It is difficult to balance the problem of robustness

Images