Pedestrian detection method based on multi-view-field graph convolutional network

Abstract

The invention discloses a pedestrian detection method based on a multi-view-field graph convolutional network. The method includes: adopting a convolutional neural network to perform feature extraction on the to-be-processed image; performing pooling and convolution processing on the extracted feature image for multiple times to obtain a preprocessed feature image; extracting a multi-scale featureinformation feature map of the preprocessed feature image by adopting a multi-view pooling pyramid; carrying out the convolution of the human body images, and obtaining a plurality of feature imagesafter convolution of the human body images; predicting and identifying a human body target area in the feature map after convolution of the human body map by adopting prediction boxes of Nbox anchor points, and completing a prediction box and a predicting category probability value. The multi-view feature pyramid is constructed by using maximum pooling, so that the detection efficiency is improved, and the problems of scale change and shielding in pedestrian detection can be effectively and efficiently solved.

Description

technical field [0001] The invention belongs to the technical field of target detection, and in particular relates to a pedestrian detection method based on a multi-view graph convolutional network. Background technique [0002] With the advent of deep learning, general object detection has made great progress, and various image processing and machine learning-based methods have been proposed to improve the performance of object detection. Although these methods show good results, considering Due to their computational cost, it is still difficult to use them in real-time systems. And there are still some limitations when it is applied to the occluded pedestrian detection task. [0003] Pedestrian detection is an important component of intelligent transportation systems that can be used to inform drivers of the location of individuals on the road for safer driving. Although some deep CNN-based methods have achieved good performance gains in general object detection, there a...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a pedestrian detection method based on multi-view graph convolutional network to overcome the scale change and occlusion problems in existing pedestrian detection

Images