基于单目视觉的动态手势轨迹识别系统分析-analysis of dynamic gesture trajectory recognition system based on monocular vision.docxVIP

优秀毕业论文 精品参考文献资料 摘要 近年来，随着计算机视觉技术的不断发展，基于视觉的动态手势识别成为了 多模式人机接口技术研究的一个重要研究方向，受到了国内外的广泛关注。本文 从手势检测和分割、动态手势跟踪和动态手势轨迹识别三个方面对基于单目视觉 的动态手势轨迹识别算法进行了研究。 在手势检测和分割部分，本文采用一种组合运动检测、肤色检测和 Gabor 纹 理检测的手势检测和分割算法。首先利用运动检测得到候选手势区域，然后对得 到的候选手势区域进行肤色分割，得到肤色候选区域。接着对得到的肤色候选区 域进行 Gabor 纹理检测。最后进行形态学处理从而得到分割效果良好的手势分割 结果。 在动态手势跟踪部分，本文使用结合 Kalman 滤波器和 Mean Shift 跟踪算法 的动态手势跟踪算法。利用 Kalman 滤波器对运动手势的位置和跟踪窗口宽度进 行预测，预测结果作为 Mean Shift 跟踪算法的输入进行运动手势的跟踪，达到了 实时跟踪的目的。 在动态手势轨迹识别部分，本文仔细研究了隐马尔可夫模型，并实现了基于 隐马尔可夫模型的手势模型训练算法和手势轨迹识别算法。 最后在以上算法研究的基础上，本文设计并实现了一个基于视频流的实时动 态手势轨迹识别系统。实验结果表明，系统具有较高识别率。 关键词： 动态手势轨迹识别 Gabor 小波变换 Kalman 滤波器 Mean Shift 隐马尔可夫模型   Abstract In recent years, with the development of computer vision technology, the dynamic gesture recognition based on vision has become one of the important research directions of the multiple pattern man-machine interface technology and received wide attention of home and abroad. This thesis presents a dynamic gesture track recognition algorithm based on monocular vision from points of gesture detection and segmentation, dynamic gesture tracking and dynamic gesture track recognition. In the part of gesture detection and segmentation, the thesis use a method, which is the combination of motion detection, skin detection and Gabor texture detection. First, we use the motion detection to locate the candidate gesture region, and then we use skin detection in the candidate gesture region to get the area of candidate skin. After that, we use Gabor texture detection in the candidate skin region. Finally, to get the better results of gesture detection and segmentation, we take morphology processing with the segmentation results. In the part of dynamic gesture tracking, the algorithm, which is the combination of Kalman filter and Mean Shift tracking algorithm, is used in this paper. Using the Kalman filter, this algorithm can forecast the position of the dynamic gesture and the width of the tracking window. To achieve the purpose of real-time tracking, we put the pr