Traffic sign and signal lamp detection and identification optimization algorithm

Abstract

The invention discloses a traffic sign and signal lamp detection and identification optimization algorithm. The algorithm is divided into a feature extraction stage, a region candidate stage and a hierarchical classification stage. In the feature extraction stage, a Ghost bottleneck module is introduced to construct a feature extraction network, big data training is carried out on ImageNet to obtain a pre-training model, image features are extracted through the pre-trained feature extraction network, and pooling processing is carried out on a feature map; in the region candidate stage, RPN subnets are adopted to obtain candidate regions, and feature maps corresponding to the candidate regions are cut and scaled, so that the sizes of the feature submaps to be classified are the same; in thehierarchical classification stage, images are classified through a hierarchical classification method, and traffic lights and traffic signs are identified. Compared with a baseline algorithm, the performance of the algorithm is greatly improved in all aspects, and the real-time performance and reliability requirements of an automatic driving system can be met.

Description

technical field [0001] The invention belongs to the technical field of statistical pattern recognition and image processing, and in particular relates to an optimization algorithm for detection and recognition of traffic signs and signal lights. Background technique [0002] Traffic sign and signal light detection is an important part of road scene perception for autonomous vehicles. The key issue in the detection of traffic signs and signal lights is to locate and identify targets while meeting real-time requirements. Specifically, it refers to checking whether the target object exists in a complex and fast-moving image sequence, and accurately and quickly calculating the target A technology of location in an image, the main problem to be solved is target recognition and positioning under complex lighting, complex background, multi-scale, multi-view, occlusion and other conditions. [0003] For the recognition of traffic lights and traffic signs, most current research rega...

AI Technical Summary

Problems solved by technology

The traditional traffic signal recognition methods mostly rely on the intuitive features such as the color and shape of the signal light, so the applicability is poor; the deep learning method regards it as a special case of target detection, such as using the YOLO target detection algorithm to detect traffic lights. detection and identification

The development of traffic sign detection and recognition methods is relatively similar, but the calculation cost of the target detection method based on deep learning is high, and the amount of calculation cannot be reduced. GPUs with high power consumption and high computing power must be used for calculations. Due to the limited power supply of the vehicle power supply, it is difficult Meet real-time requirements

Images