Automatic Traffic Light Recognition for Mobile Robot Applications

摘要

In this study, we present a method to enhance the quality of traffic light images captured by a webcam using hardware filters for mobile robot applications. The images are processed through pre-processing, detection, and recognition steps to recognize the status of traffic lights. The results show that the status of traffic lights is recognized successfully; the overall recognition rate is about 99% while the recognition time is about 41ms to 76ms per frame. By using the method of support vector machine (SVM), the recognition rate is higher with the cost of a larger processing time.