摘要
行人检测是近年来计算机视觉领域中备受关注的前沿方向和研究热点。以单目视觉传感器作为外界环境信息获取的主要手段,建立了一个包含行人分割、识别的检测系统。根据行人特有的一些特征,提出了基于垂直边缘和边缘对称性的行人分割方法,并进行精确定位。在行人识别阶段利用HOG特征进行特征提取,然后利用线性支持向量机进行行人识别。对大量的包括不同天气和场景条件下的测试集进行了测试,结果表明:提出的算法具有良好的检测效果。
Pedestrian detection is intensively investigated and becoming a hot topic in the field of computer vision.By making use of monocular vision detector as the main mean of catching outside environmental information,a pedestrian detection system including segmenting of regions of interests（RoIs）and recognizing detection system is built.According to the particular characteristic of pedestrian,based on pedestrian segmenting method grounding on vertical edge and the symmetry property of it,and the pedestrian is accurately located and segmented from the video image.In the recognition process,HOG feature extraction method is produced to extract pedestrian features and a linear support vector machine（SVM）is used for pedestrian recognition.A large number of tests at different kinds of weather and scenes are carried out.Experimental results show that the pedestrian detection algorithm has effective performance.
出处
《传感器与微系统》
CSCD
北大核心
2011年第7期68-70,74,共4页
Transducer and Microsystem Technologies
基金
廊坊市科研计划资助项目(2010085)
关键词
行人检测
垂直边缘
边缘对称性
HOG特征
线性
支持向量机
pedestrian detection
vertical edge
symmetry property of edge
HOG feature
linear
support vector machine（SVM）
