论文部分内容阅读
针对增强现实领域手势交互的需求,提出鱼眼图像中手部定位及手势识别的方法。首先在YCbCr颜色空间中检测肤色区域，再利用HSV空间颜色直方图特征和形状特征在肤色区域中检测标识物，分割手部区域。然后将标识物质心作为手部定位中心，根据鱼眼镜头的成像原理，计算该点相对于鱼眼镜头的三维位置。最后提取手势形状特征，实现手势识别。通过对85幅不同背景及光照条件下的鱼眼图像进行实验，此方法识别手势的正确率为96%，识别单幅鱼眼图像约耗时0.03 s。实验结果表明，该方法可以快速准确地实现在复杂背景下的手部定位与手势识别。
Aiming at the need of hand gesture interaction in augmented reality field, a method of hand positioning and gesture recognition in fisheye image is proposed. Firstly, the skin color region is detected in the YCbCr color space, and then the HSV space color histogram feature and the shape feature are used to detect the marker in the skin color region to segment the hand region. Then, the centroid of the marker is used as the hand positioning center, and the three-dimensional position of the fisheye lens is calculated according to the imaging principle of the fisheye lens. Finally, extract the shape features of gestures to achieve gesture recognition. Experiments on fisheye images under 85 different background and light conditions showed that the accuracy of this method was 96% and the recognition time of single fisheye image was 0.03 s. Experimental results show that this method can quickly and accurately hand position and gesture recognition in complex background.