%0 Journal Article
%T 采用超像素标注匹配的交通场景几何分割方法<br>A Geometric Segmentation Method for Traffic Scenes Using Super？？Pixel Label Matching
%A 胡江策
%A 卢朝阳
%A 李静
%A 邓燕子
%A 刘阳
%J 西安交通大学学报
%D 2018
%R 10.7652/xjtuxb201808012
%X 针对交通场景逐像素标注方法计算复杂、模型训练耗时长的问题,提出了一种基于超像素标注匹配的交通场景几何分割方法。该方法无需进行模型训练,根据全局特征搜索一组待分割交通场景图像的相似图像集。对待分割图像进行超像素分割和超像素块特征提取,并利用朴素贝叶斯原理进行似然比计算,根据似然比在相似图像集中进行超像素块标注匹配以实现初次分割。利用初次分割结果计算出一元势,应用全连接条件随机场模型对初次分割结果进行优化。实验结果表明,与传统的逐像素标注方法相比,本文方法的分割正确率和平均召回率分别提高了4％和3％,能够有效地实现交通场景几何分割。<br>A geometric segmentation method for traffic scenes based on super？？pixel label matching is proposed to solve the problem of complex calculation and long time？？consumption of model training in the pixel？？by？？pixel labeling method for traffic scenes. The proposed method does not require model training, and a set of images similar to picture a traffic scene that will be segmented is searched according to global features. Then, super？？pixel segmentation and super？？pixel block feature extraction are performed on the traffic scene, and the likelihood ratio is calculated using the naive Bayesian principle. Super？？pixel block label matching is performed in the set of similar images according to the likelihood ratio to realize an initial segmentation. Finally, the initial segmentation result is used to calculate the unary potential, and the fully connected conditional random field model is used to optimize the initial segmentation result. Experimental results and a comparison with the traditional pixel？？by？？pixel labeling method show that the proposed method effectively achieves geometric segmentation of traffic scenes, and the accuracy of segmentation and the average recall rate increase by 4％ and 3％ respectively
%K 交通场景
%K 超像素
%K 几何分割
%K 全连接条件随机场&lt
%K br&gt
%K traffic scene
%K super？？pixel
%K geometric segmentation
%K fully connected conditional random field
%U http://zkxb.xjtu.edu.cn/oa/DArticle.aspx?type=view&id=201808012