Research on Image Recognition Algorithm of Unmanned Vehicle Based on Fast Running

摘要

This paper proposes an efficient method for identifying specific traffic signs for fast-running unmanned cars, and solves the problem of visual navigation for unmanned vehicles under fast operating conditions. Firstly, the traffic signs with different shapes are classified by the Minimum Area Bounding Rectangle (MABR) algorithm, and each specific identification image is analyzed in combination with the Histogram of Oriented Gradient (HOG) feature so as to quickly identify different Shape traffic sign. By setting up a fast-running unmanned vehicle operating platform, the real-time, accuracy, and robustness of the design method are verified.