论文部分内容阅读
针对复杂背景下红外微弱点状运动目标的检测,提出一种基于小波域HMT模型的图像杂波抑制方法。对图像小波系数低频部分建立隐马尔可夫树模型,使用Bayesian准则估计图像背景小波系数,参照杂波抑制模型,得到杂波抑制后图像的信号加噪声模型,并通过计算Kendall秩相关系数和Friedman统计量验证了该方法残留噪声的高斯性和独立性。
Aiming at the detection of infra-red weak punctuated moving targets under complicated background, an image clutter suppression method based on wavelet domain HMT model is proposed. The hidden Markov tree model is established for the low frequency part of the image wavelet coefficients. The Bayesian criterion is used to estimate the image background wavelet coefficients. The clutter suppression model is used to obtain the signal plus noise model of clutter-suppressed image. By calculating the Kendall rank correlation coefficient and Friedman Statistics verify the Gaussianity and independence of the residual noise of this method.