基于内侧轮廓模型的粒子滤波轮廓跟踪

为了提高复杂环境下轮廓跟踪的鲁棒性，提出了一种基于内侧轮廓模型的粒子滤波轮廓跟踪算法.① 在轮廓采样点的法线上利用Canny算子得到轮廓法线方向的梯度信息；② 用该法线的内侧部分构建局部颜色信息，并使之与梯度信息结合，形成一个新的一维法线观测似然；③ 用所有内侧法线构建一幅全局内侧颜色直方图；④ 将梯度信息、局部颜色信息和全局颜色信息3种特征进行有效融合，形成一个新的多特征融合观测模型.实验结果表明，在复杂环境下，该算法能够较好地实现对复杂的非封闭轮廓的鲁棒跟踪.

Visual Contour Tracking Based on Inner-contour Particle Filter

This paper presented a novel particle filter called Inner-Contour Particle Filter to track unclosed contour under complex background. Aimed at achieving effectiveness and robustness against complex background, the proposed algorithm first used Canny edge detector to detect the edge information along the normal line of the contour, and then sampled the inner part of the normal line to get the local color information and combined it with the edge information to construct a new normal line likehood. After that, all the inner color information was used to construct a global color histogram. Finally, the edge information, local color information and global color information were integrated to form a new conservation likehood. Experimental results show that the proposed method is robust and computationally efficient for tracking contours under complex background, and can run in real time completely.