Maneuvering Target Tracking Algorithm Based on Muti-paramter Sequential Extended Kalman Filter

Based on the information theory,the performance of maneuvering target tracking can be improved by increasing the input information( observation vector).In this paper,the estimations of radial acceleration and radial velocity obtained in the signal processing are introduced into the measurement vector by coordinate transformation.In order to solve the problem of high nonlinearity of the radial acceleration,radial velocity and the state vector,a new algorithm of multi-parameter sequential extended Kalman filter( MSEKF) is proposed.The tracking performance of this algorithm is tested and compared with the other tracking algorithms.It is shown that the proposed algorithm outperforms these algorithms in strong and weak maneuvering environments.     

基于局部像素特征分类的图像分割算法

图像分割是模式识别、计算机视觉等领域的重要研究内容,也是图像信息处理的难点和热点之一.以孪生支持向量机(TSVM)与极坐标复指数变换(PCET)理论为基础,提出了一种基于局部像素特征分类的图像分割算法.该算法首先对局部像素窗口进行PCET,并将PCET系数幅值作为图像的像素级特征;然后利用指数交叉熵阈值技术确定训练样本,并进一步训练出TSVM分类模型;最后利用已获得的TSVM分类模型对原图像像素进行分类处理,从而获得图像的最终分割结果.实验结果表明,该算法可以获得较好的图像分割结果.