多源遥感数字图像复原优化研究

摘要

A restoration method of digital image of multi-source remote sensing is proposed based on nonlinear least squares.The extreme point of scale space of digital image integrated with SIFT algorithm is detected and contrast ratio of fuzzy feature point of image is worked out,then the feature of gradient direction distribution of neighborhood pixel is used as provided direction parameter of each image feature point.The initial matching pair of image integrated with correlation matching theory is obtained,and the matching strength of image feature is worked out.Thus,the digital image restoration is completed.Experimental results show that the proposed method can improve the effect of image restoration.Reliability of digital image interpretation is excellent.%对多源遥感数字图像进行复原,可有效解决图像部分边缘丢失和纹理信息模糊的问题.对数字图像的复原,需要获得图像初始匹配对,计算出图像特征匹配强度,完成数字图像的复原.传统方法通过组建数字图像模糊近似度,构造模糊特征加权因子,但无法得出特征匹配强度,导致复原效果不理想.提出基于非线性最小二乘的多源遥感数字图像复原方法.结合SIFT算子检测多源遥感数字图像尺度空间极值点,计算出图像模糊特征点对比度值,将图像模糊特征点邻域像素的梯度方向分布特性作为各个图像模糊特征点指定方向参数,结合相关匹配理论获得图像初始匹配对,计算出图像特征匹配强度, 以此为依据完成多源遥感数字图像复原.实验结果表明,所提方法能够有效提升多源遥感数字图像复原效果,合成多源遥感数字图像判读的可靠性较优.