中文摘要：

传统的目标检测方法需要对大量候选窗（区域）做判断,需要较大的计算量。本文根据人体特点,提出了一种基于分级判断的方法,需要判断的候选窗逐级减少,因此可以大量减少复杂特征和分类器需要判断的候选窗数量,进而减少整个检测算法的计算量。算法首先对待检测图像提取NG（norm of gradients）特征,通过线性支持向量机（SVM）判断得到行人的候选区域;然后对候选区域提取简化梯度方向直方图（HOG,histograms of oriented gradients）特征,采用线性SVM对候选区域进一步的过滤;最后对经过过滤筛选得到的区域提取多分辨率HOG特征,使用可变形部件模型（DPM,deformation part model）对候选区域进行检测定位行人的位置。在INRIA数据集上的实验结果表明,本文方法在保证检测精度的情况下,虽然相比于原始DPM算法有少数的行人漏检,但是本文方法的检测结果中行人误检数目远少于原始DPM算法,检测速度也优于原始DPM算法。

英文摘要：

Traditional object detection method needs to judge a large number of candidate windows（regions）,so it needs a large amount of calculation.In this paper,according to pedestrian characteristics,we put forward a method based on hierarchical judgment.The candidate windows that need to be detected is reduced progressively,so we can reduce a large number of candidate regions that need to be judged by complicated feature and classifier,which reduces the amount of calculation of the whole algorithm.Firstly,we extract norm of the gradients（NG）feature from image,and use linear support vector machines（SVM）to get the candidate regions of the pedestrians.Secondly,we extract simple Histograms of oriented gradients（HOG）feature from the candidate regions,and the candidate regions are further filtered by linear SVM.Finally,we extract multi-resolution HOG feature from the filtered candidate regions,deformation part model（DPM）to detect the candidate areas to locate the precise location of pedestrians.On the INRIA dataset,experimental results show that on the basic of ensuring the accuracy of detection,a small number of pedestrian are not detected compared with the original DPM algorithm,but the number of error detection is far less than that of the original DPM algorithm,and the detection speed is faster than that of the original DPM algorithm.