基于Kinect骨骼信息和改进重心距离法的指尖识别

Detection of Fingertip with Improved Focus Distance Method Based on Kinect Skeleton Tracking

针对肤色识别易受环境影响问题,提出了一种基于Kinect骨骼信息和改进重心距离法的手指指尖识别方法。首先采用微软Kinect摄像头获取人体骨骼信息,并将双手坐标点的三维信息转换成彩色图上的二维信息进行手掌区域的提取。然后利用OpenCV基于肤色检测算法将手掌轮廓从背景图像中分割出来。最后针对重心距离法鲁棒性差的特点提出一个改进因子。实验结果证明,该方法可以高效地检测出手指的指尖位置,识别出人体的手指。实验结果表明该方法能有效排除类肤色区域和手指轮廓不足对指尖检测造成的影响,具有较高的检测精度。

For color identifying of fingertips is vulnerable to environmental impacts. A hand contour extraction method and fin-gertip recognition method based on skeleton data was presented. Firstly, Microsoft Kinect camera obtain information on the hu-man skeleton , converted three-dimensional information of hand skeleton point into two-dimension on RGB map, the position of hand was found quickly .Then, use OpenCV detection algorithm based on color palm information extraction regional palm from the background image. Finally, the robustness of the characteristics of poor focus distance method presents an improved fac-tor Experimental results show that the method can detect the contour of straight and curved fingers successfully and it can be used in a dark environment.