Extraction of Lunar Surface Geological Features with Chang'e-1 Data Using Coherent Algorithm
-
摘要: 经典的边缘提取算法可从月球影像数据中提取出月表的构造特征, 但难以提取出与月貌有关的特征信息.本文提出的相干法月球影像地质特征提取技术, 其基本思想是基于相干运算来考察图像上像素或像素组之间的相似性, 以获取月表的构造特征和月貌特征.该方法已经应用在嫦娥一号月球探测器获取的部分影像数据上, 相比于几种常规的边缘提取方法, 本方法优势明显, 不仅能有效地提取环形构造和线性构造等地质特征, 还可以很好地提取月貌特征, 对具有一定规模的小直径撞击坑的识别也有不错的效果.该方法可以用于计算机自动识别或辅助手工解译月表构造, 划分月貌单元等研究.Abstract: The classical edge extracting algorithms in the image processing can extract lunar surface geological features, but can not extract texture features of lunar terrain from Chang'e-1 image data. This paper suggests a method of coherent algorithm to extract the lunar surface geological features with Chang'e-1 data, which is based on coherent calculation of signals in digital signal processing transplanted into image processing for lunar data. We processed some Chang'e-1 image data by using coherent calculation which involves calculating the similarity between pixels or pixel groups. By comparing several classical algorithms with our algorithm, the effectiveness of this method is demonstrated. The coherent algorithm can extract linear and circular geological structure features as well as the texture features of lunar terrain from Chang'e-1 lunar image at the same time. And it is also able to identify the small-diameter craters. So the method can be applied to computer-driven automatic recognition of lunar structures and landforms for evolution analysis of the moon.
-
Key words:
- lunar surface /
- edge extraction /
- geological structure /
- lunar landform /
- coherent algorithm /
- geophysics
-
图 1 X和Y像素矩阵的相干运算示意图(符号⊙表示相干计算)
Fig. 1. Illustration of coherent calculation between X and Y pixel matrices
图 3 月球影像不同相干参数处理结果示例
a.原始灰度图(嫦娥拍摄的影像数据，经度71.112 3°~77.821 8°E，纬度74.126 9°~74.126 3°S，分辨率120 m)；b.M=2，P=0，Q=1处理结果；c.M=3，P=0，Q=1处理结果；d.M=3，P=1，Q=1处理结果；e.M=3，P=0，Q=3处理结果；f.M=10，P=0，Q=1处理结果
Fig. 3. Illustration of different coherent parameters with the lunar image data
图 4 月球影像不同边缘提取算法处理结果
a.原始灰度图(嫦娥拍摄的影像数据，经度71.112 3°~77.821 8°E，纬度74.134 4°~74.133 8°S，分辨率120 m)；b.用Gauss-Laplace法处理的结果；c.用Robert法处理的结果；d.用相干算法处理的结果
Fig. 4. Illustration of different edge-extracted algorithms with the lunar image data
表 1 用于相干试验的参数
Table 1. Parameters for coherent algorithm test
图幅编号 参数 矩阵 M P Q b 2 0 1 2×2 c 3 0 1 3×3 d 3 1 1 3×3 e 3 0 3 3×3 f 10 0 1 10×10 
下载: 导出CSV
表 2 相干算法的参数影响
Table 2. Effect of coherent algorithm under various parameters
参数 M P Q 取值范围 1~width - M~+ M - M~+ M 变换趋势 大 小 绝对值从大到小 绝对值从大到小 对图像的影响 结构突出 边缘清晰 纵向分辨率逐渐提高 横向分辨率逐渐提高 对地质特征的影响 可提取大构造特征 可提取小构造特征 纹理特征逐渐清晰 纹理特征逐渐清晰 注：width为原始图像宽度.
下载: 导出CSV
-
[1] Dragomir, S.S., 2000. On the Cauchy-Buniakowsky-Schwartz's inequality for sequences in inner product spaces. Mathematical Inequalities and Applications, 3(3): 385-398. http://www.researchgate.net/profile/S_Dragomir/publication/255667006_On_the_Cauchy-Buniakowsky-Schwartz%27s_inequality_for_sequences_in_inner_product_spaces/links/0046353bb7e9b3c378000000/On-the-Cauchy-Buniakowsky-Schwartzs-inequality-for-sequences-in-inner-product-spaces.pdf [2] Hu, G.S., 2005. Introduction of digital signal processing. Tsinghua University Press, Beijing, 26-31 (in Chinese). [3] Ouyang, Z.Y., 2005. The introduction of lunar science. China Astronautic Press, Beijing, 1-200 (in Chinese). [4] Wang, H., 2007. The analysis of lunar surface features based on remote sensing image (Dissertation). Jilin University, Changchun, 1-45 (in Chinese). [5] Wu, W.R., 2008. To the moon. China Astronautic Press, Beijing, 85-122 (in Chinese). [6] Zhang, H.Y., Wu, B., Peng, Q.C., 2005. An improved algorithm for image edge detection based on lifting scheme. Journal of Electronic Science and Technology of China, 3(2): 113-116. http://epub.cnki.net/grid2008/docdown/docdownload.aspx?filename=ZGKE200502004&dbcode=CJFD&year=2005&dflag=pdfdown [7] 胡广书, 2005. 数字信号处理导论. 北京: 清华大学出版社, 26-31. [8] 欧阳自远, 2005. 月球科学概论. 北京: 中国宇航出版社, 1-200. [9] 王洪, 2007. 基于遥感图像的月球表面特征分析(学位论文). 长春: 吉林大学, 1-45. [10] 吴伟仁, 2008. 奔向月球. 北京: 中国宇航出版社, 85-122. -
点击查看大图
计量
- 文章访问数: 3333
- HTML全文浏览量: 126
- PDF下载量: 55
- 被引次数: 0
