论文部分内容阅读
撞击坑是月球表面最重要的地质构造之一,通过对“嫦娥一号”CCD影像中撞击坑的边缘清晰度进行评价,可以进一步反演出月球表面的风化程度、地表起伏等地质信息。提出一种基于图像清晰度评价的边缘清晰度评价方法,从空域的梯度、频域的高频分量以及信息论三个方面,运用基于Sobel算子、小波变换和信息熵的算法对撞击坑的边缘清晰度予以评价。设计出一种适应于月球撞击坑特征的BP神经网络,组合三种评价算法的结果作为其输入,进而得到最终的清晰度等级。将最终结果加载到具有自主知识产权的数字月球平台上予以全月性的展示和进一步分析。
Crash pits are one of the most important geological structures on the lunar surface. By evaluating the sharpness of the craters in the “Chang’e-1” CCD imagery, we can further derive the weathering degree and surface geological information of the lunar surface. This paper proposes a method for evaluating the sharpness of the edge based on image sharpness evaluation. From the aspects of airspace gradient, high frequency component in the frequency domain and information theory, this paper uses Sobel operator, wavelet transform and information entropy algorithm to evaluate the edge of impact crater Clarity to be evaluated. A BP neural network adapted to the characteristics of lunar craters is designed and the results of the three evaluation algorithms are combined as input to obtain the final definition level. The final result is loaded onto the digital lunar platform with independent intellectual property rights for full-time display and further analysis.