Pedestrian detection based on HOG
no vote
Application background pedestrian detection for OpenCV, Kinect,, HOG,, pedestrain,, is a kind of pedestrian detection (detection).. The program contains two segments of code, using HOG OpenCV pedestrian detection class to complete the pedestrian detection in the picture sequence. Two different sizes for the detection of different picture effect is not the same. For the MIT pedestrian library picture is small, with second good; for the detection of the big picture is the first good. Key Technology & nbsp; & nbsp; & nbsp; & nbsp; hog feature descriptor (hog descriptors) is by navneet DALAL and bill Triggs referred to in 2005 of a pedestrian detection method of the feature descriptor. The main idea is to calculate the statistical value of the directional information of the local image gradient, as the local eigenvalues of the image.. The main steps are:After the normalized images, because we do not help color data, so the