Traffic light recognition algorithm based on convolutional network

Abstract

Disclosed is a traffic light recognition algorithm based on a convolutional network, comprising an original image acquisition and processing module, an input image feature extraction module, a candidate area acquisition module, a traffic light candidate area acquisition module, a traffic light detection module, a color and shape recognition module and a recognition result output module.

Description

technical field [0001] The invention relates to the field of environment perception of intelligent vehicles and the field of traffic safety, in particular to a traffic signal light recognition algorithm based on a convolutional network. Background technique [0002] In recent years, the number of civilian cars in the country has continued to increase, but many drivers have not developed good driving habits, which leads to frequent traffic accidents. Therefore, improving the safety of traffic operation has become a hot issue in the society, and research on vehicle-mounted assisted driving systems has also become a hot topic. [0003] In the vehicle assisted driving system, traffic signal light recognition is particularly important. It can not only assist the driver in driving, but also help color-blind and color-weak people understand the status of road traffic lights ahead. However, traffic signal recognition is affected by distance, resulting in different scales, and it is...

AI Technical Summary

Problems solved by technology

[0007] This invention requires a standard black background frame message during daylight hours, but traffic light background panels are not standard in many areas

Images