基于线性SVM的车辆前方行人检测方法

Method of pedestrian detection ahead of vehicle based on linear SVM

采用基于线性SVM方法检测复杂交通背景下车辆前方行人.该方法根据行人非刚性的特点,利用三线性插值法提取图像的梯度方向直方图特征,采用线性支持向量机对视频中的图像进行多尺度融合检测,以适应复杂交通背景的行人检测需求,有效提高检测准确性.实验表明,该算法能够对混合交通视频中的不同尺度和姿态的行人进行有效识别.

The pedestrians ahead of vehicle in complex transportation background were detected by using the pedestrian detection method based on linear SVM. According to the characteristics of pedestrian non-rigid, trilinear interpolation algorithm was used to extract the histograms of oriented gradient features of images. Linear support vector machine （SVM） was used to can＇y on the multi-scale fusion examination of the video image in order to adapt to the requirement of pedestrian detection in complex transportation background and to effectively improve the detection accuracy. The experiment results show that the algorithm can effectively identify the different scales and poses pedestrian in mixed traffic videos.