一种新型月球地形自动识别迭代算法

发布时间：2018-01-05 10:42

本文关键词：一种新型月球地形自动识别迭代算法　出处：《宇航学报》2017年01期 　论文类型：期刊论文

【摘要】：针对现有月球地形自动识别算法的识别率和精确度较低的问题,提出一种结合CCD图像和DEM数据信息,自动识别月球地形的动态分块迭代算法,实现了识别率和识别精度的双重提高。新算法提取CCD和DEM数据中月表地形的不同特征来构建图像子块的特征向量,再对特征向量聚类区分月球地形。算法根据输入图像的精度决定初始子块尺寸,提取子块的特征向量后聚类区分月海、月陆。每轮输出分类可信度高的子块结果后,会对分类结果可信度较低的子块进行细分,对细分后的子块重新提取特征向量并再次聚类分类,直到迭代算法终止。新算法已在三个典型的月面区域:虹湾(SI)、H010和危海(Crisium)区域进行了测试,试验结果与现有的地形分块识别算法相比,新算法的识别率和相关kappa系数均优于已知的自动识别算法结果。
[Abstract]:In order to solve the problem of low recognition rate and accuracy of the existing automatic lunar terrain recognition algorithms, a dynamic block iterative algorithm based on CCD images and DEM data is proposed. The new algorithm extracts the different features of the monthly table terrain in CCD and DEM data to construct the feature vector of the image sub-block. The algorithm determines the size of the initial sub-block according to the accuracy of the input image and extracts the feature vector of the sub-block to distinguish the lunar sea. After each round output the sub-block with high classification credibility, the sub-block with low credibility will be subdivided, and the subblock after subdivision will be extracted the feature vector again and clustered again. Until the termination of the iterative algorithm, the new algorithm has been tested in three typical lunar areas: siphon H010 and Crisium. The experimental results show that the recognition rate and the correlation kappa coefficient of the new algorithm are better than those of the known automatic recognition algorithm.
【作者单位】：澳门科技大学资讯科技学院;澳门科技大学月球与行星科学实验室/太空科学研究所;
【基金】：澳门科学技术发展基金(059/2013/A2,039/2013/A2)
【分类号】：TP751
【正文快照】： 0引言人类对太空天体的探测已历经多年,其中月球探测长期以来备受各国科学家的关注[1]。月球探测起源于17世纪,著名天文学家伽利略将他创制的天文望远镜对准了月球[2]。近年来,随着航天技术的不断发展,人类所获得的月球数据逐渐增多,有关月球的研究项目成果也是层出不穷。虽然

【相似文献】