基于线矩特征的破损目标识别

基于线矩特征的破损目标识别 RECOGNITION OF OCCLUDED OBJECTS WITH LINE MOMENTS 赵恒卓 Zhao Hengzhuo 1 first-author 杨春旭 Yang Chunxu 1 胡正仪 Hu Zhengyi 1 王延平 Wang Yanping 1 武汉大学电子信息学学院 武汉大学电子信息学学院 College of Electronic Information,Wuhan University College of Electronic Information,Wuhan University 论述了用曲线的矩特征作为局部不变量来识别二维破损目标的方法．该方法对目标的轮廓进行分割后，利用阶次规正后的轮廓线矩特征，通过粗、细两级的基于Ｈｏｕｇｈ变换的投票匹配法来识别目标．分类实验表明，该方法有利于提高目标局部特征的抗噪性，并简化了破损目标的匹配和识别算法． This paper presents a method for recognizing 2D occluded objects with line moments. By partitioning the image contour based on the breakpoints’ positions and using the line moments as the local features of the image contour, this methord can recognize the occluded objects in a two gradational matching procedures which are similar to Hough transformation. The experiment shows that the proposed method effectively improves the features stability to noise and simplifies the matching procedures of the objects. 破损目标 不变矩 阶次规正 线矩 拐点 Hough变换 occuluded object Invariant moment order normalization line moment corner Hough tansformation TN919.8 国防科技预研基金 1997-05-01 2021-04-01 5