In order to solve the problem of low r ecognition rate and time complexity of traffic sign recognition,this paper proposes an improved HOG feature extraction and extreme learning machine ( ELM) traffic sign recognition algorithm.By describing the direction gradient histogram (HOG) feature of the e dge information of the traffic sign image,and integrating with the local binary pattern (LBP) feature capable of re presenting the internal texture information of the logo image,a HOGv-LBP effective feature is formed and then utilized.ELM conducts traffic sign training and classification.The experimental results show that the proposed alg orithm not only improves the recognition rate of traffic signs,but also reduces the time complexity and enha nces the system robustness.