中文摘要：

针对复杂图像分割问题开展研究,并以机器人视觉中目标搜索和识别问题为支撑目标,结合该背景明确提出了图像分割算法性能评价标准和侧重点,基于此约束,以Mean Shift分割方法为基础,并重点考虑了分割尺度的有效控制、分割过程兼顾场景深度信息等问题,对算法进行了针对性改进。针对分割尺度控制问题,提出了边缘敏感度的概念,提高了算法尺度分块的控制能力。针对深度信息融合问题,采用了双目视觉立体匹配和基于Kinect传感器的两种深度信息获取方法,均成功实现融合并提高了分割效果。实验结果表明,本文算法与传统Mean Shift算法相比具有明显优势,不仅能更有效地控制分割尺度,还能成功分割原算法难以分割的特殊情况。

英文摘要：

Complex image segmentation is studied to solve the problem of target searching and identification in robot vision. In this context, performance evaluation criteria and emphasis for image segmentation algorithm are proposed explicitly. A modified algorithm is developed based on Mean Shift segmentation methods fully taking into account the effective control of segmentation scale and the scene depth information of the segmenta-tion process. In order to solve the problem of controlling segmentation scale, the concept of edge sensitivity is proposed so that the control ability of the algorithm scale block is improved. Two methods of getting the depth information based on binocular stereo matching and image sensor Kinect are employed to deal with the depth in-formation fusion. Both methods realize the fusion successfully and the performances of segmentation are en-hanced obviously. The experiment results demonstrate that the proposed method is more preponderant than traditional algorithms. The proposed algorithm could not only control segmentation scale, but also split the special situations.