本文针对高速环境下的车型识别问题, 提出基于方向可控滤波器的改进HOG算法。将方向可控滤波器算法与HOG算法相结合, 以实现对车辆图像特征提取。采用主成分分析算法（PCA)约减特征向量维数以减少计算复杂度, 利用支持向量机算法对提取特征进行样本训练, 实现对车辆外型特征的识别。仿真实验结果表明: 采用该算法原始车辆车型的识别正确率均值达到9236%; 另外, 本文方法的识别速度比传统的HOG特征算法提高了345%, 识别实时性得到提升。本文算法比传统HOG算法更优, 能有效提高车型识别的效率。

Aiming at problems of vehicle type recognition in high-speed environment, an improved HOG algorithm based on oriented steerable filter is proposed in this paper. Vehicle image features are extracted by combining the oriented steerable filter algorithm and HOG algorithm. The principle component analysis(PCA) is used to reduce dimensions of the eigenvector for decreasing the computational complexity. The support vector machine(SVM) algorithm is used to train the extracted features to realize the recognition of vehicle′s appearance features. The simulation results indicate that average vehicle type recognition accurate of proposed algorithm reaches 9236%. At the same time, the recognition speed is 345% higher than the traditional HOG feature algorithm. In conclusion, the proposed algorithm can effectively improve the efficiency of vehicle type recognition and is therefore better than the traditional HOG algorithm.