协方差矩阵在目标二维姿态测量中的应用

针对现有姿态测量方法计算量大的问题，本文提出了一种新的基于协方差矩阵的目标二维姿态测量方法。首先，将图像二值化，提取出目标的重心坐标；其次，计算出目标所在区域的协方差矩阵；最后，得到由协方差矩阵定义的椭圆区域，并求出椭圆的两个轴的大小和方向。两幅图像中目标区域所确定的椭圆长(短)轴的比值即为缩放比例，长(短)轴方向的差值即为旋转角度。仿真实验证明：该方法可以快速准确地计算出目标的旋转角度和缩放比例，旋转角度可以精确到 以内，缩放比例可以精确到 以内，计算时间仅为对数极坐标的1/16。

Application of covariance matrix to 2D object attitude measurement

Aiming at large amount of computation, a new method of 2-D object attitude measurement based on covariance matrix is presented. Firstly, binary the image and calculate the center of object. Then, compute covariance matrix of the object area. Finally, get the ellipse area defined by covariance matrix, and compute the length and orientation of both axes. The ratio of corresponding axes is the scale change, and the orientation difference of corresponding axes is the rotation angle. The experimental results show that the presented method can calculate the rotation and scale change fast and accurately, the angle error is within , and the scale error is within ,while the computation time is 1/16 of that of log-polar method.