基于联合标定的图像分割

针对普适环境下人体区域的提取问题,提出一种基于彩色摄像头和深度摄像头联合标定的图像分割方法。首先根据深度摄像头采集的深度信息易处理和受外界环境影响小的优点,利用大律法从深度图像中快速提取出人体区域;然后建立世界坐标系中彩色图像坐标系和深度图像坐标系的对应关系,实现彩色信息和深度信息的相对位置映射;最后根据分割的深度图像及其在彩色图像中的映射,在彩色图像中完成人体区域分割。通过在复杂背景和不同光照条件下的图像分割实验,验证了本方法具有良好的精度和鲁棒性。

Image Segmentation Based on Joint Calibration

Concerning the problem that it’s dif-ficult to extract the body region under pervasive environment,we proposed an image segmentation method based on joint calibration of color camera and depth camera.In a complex environment in-doors or outdoors,as a depth camera can capture information fast and won’t be affected by the envi-ronment,we use OTSU to extract body region from a depth image.Then we use the joint calibra-tion method proposed in this article to get the body region from the color image.At last,by experi-ments of different backgrounds and illumination conditions,the validity of the method is verified.In this way,our method eliminates the effect of light and complex background on the extraction.Also,our method can extract dynamic or static body re-gion quickly and accurately.Our method is robust and real time.