利用K-L变换进行人脸识别。首先求得待辨识图像相对于训练集平均脸的差值图像，然后求得该图像在特征脸空间中的坐标，最后采用最近邻法对图像进行归类。-KL transform for face recognition. Obtain the first image to be identified image with respect to the difference between the average face of the training set, and then obtain the

坐标变换矩阵，将机械手的相对位姿变换为绝对位姿，类似于欧拉转角矩阵的一个小程序-Coordinate transformation matrix, the relative position and orientation of the robot is transformed absolute posture, similar to a small program Euler angle matrix