论文部分内容阅读
含有点状运动目标的红外图像序列情况下,在研究杂波运动机理和数理模型的基础上,假定杂波是时空非平稳随机过程,采用时空非参数预测(时域衰减正弦、空域Gabor加权函数)技术来抑制了动态背景杂波。另外,杂波抑制后,残留噪声的高斯性和独立性通过Kendall秩相关法和计算Friedman统计量的方法来进行了验证,其结果表明本文所研究的算法思路的可行性。
In the case of infrared image sequences containing point moving targets, based on studying clutter motion mechanism and mathematical model, it is assumed that clutter is a spatio-temporal nonstationary stochastic process. The spatio-temporal non-parametric prediction (time-domain decay sine, space Gabor weighting function ) Technology to suppress dynamic background clutter. In addition, after clutter suppression, the Gaussianity and independence of residual noise are verified by Kendall rank correlation method and Friedman’s statistic method. The results show the feasibility of the algorithm ideas studied in this paper.