An Optimization Algorithm of Intelligent Traffic Image Recognition Technology

Abstract

The implementation of traffic sign recognition is of great significance for promoting social progress and the harmony of human environment. The purpose of this paper is to study the optimization algorithm of intelligent image recognition technology. In order to reduce the total number of parameters in the training process, reduce the memory capacity, but ensure the accuracy of network identification, based on the AlexNet model, two improvement and optimization dimensions are proposed for design and network development. AlexNet deep network and architecture network, defining model parameters. The consistency of the model and the progress of the algorithm are demonstrated by observing the normal training method and the performance loss with the number of repetitions. The final application of traffic sign recognition achieved good results, achieving accurate recognition of 90% of the GTSRB data. The traffic recognition algorithm in this paper is trained on data sets from real scenes, so that the algorithm has refined and powerful capabilities for the isolation and identification of practical application scenarios, and has a good automatic algorithm.