Collaborative Research: A Statistical Learning Tool for the Analysis and Characterization of Mars Topography

该项目的目标是设计和开发一种用于对火星地形特征进行分类和定性的空间学习工具。 研究火星表面的主要工具是地貌测绘和地质测绘。执行这些映射的标准方法是通过手动解释图像。这种费力的方法严重限制了适合研究的火星站点的数量。STL实现了地貌填图的自动化，加快了地质填图的速度。 因此，它能够快速和定量表征火星表面的大部分。火星探测器使用数字地形而不是图像来描述火星遗址。不同的地形变量融合成一个多层的数据结构。一个地点的每个像素都携带着一系列当地和区域的地形信息。地形特征的自动识别和分类是在像元级上进行的。这使得定量表征和比较不同的地形形成的基础上，其组成像素的统计。利用地形专题图可以方便地将结果可视化。通过增加其他数据类型（多光谱图像）并将其应用于其他行星表面，可以扩大卫星图像的能力。这一方法有可能成为一个强大的调查工具，具有广泛的应用。为了便于研究界采用，将把实现该工具的代码及其文档放在公共领域。这项工作的成果将通过新的课程、研讨会和与其他机构的合作加以传播。

The goal of this project is to design and develop a statistical-learning tool (STL) for classification and characterization of topographical features on Mars. Major tools for studying the Martian surface are geomorphic mapping and geologic mapping. The standard approach to perform these mappings is through a manual interpretation of images. This laborious approach severely limits the number of Martian sites amenable to study. The STL automates geomorphic mapping and expedites geologic mapping. Thus, it enables fast and quantitative characterization of large sections of the Martian surface. The SLT uses digital topography instead of images to characterize Martian sites. Different topographical variables are fused into a multi-layer data structure. Each pixel in a site carries an array of local and regional topographic information. The automatic recognition and classification of topographic features is performed at the pixel level. This enables the quantitative characterization and comparison of different topographic formations based on statistics of their constituent pixels. The results can be conveniently visualized by means of thematic maps of topography. The capacity of the SLT can be extended by adding other data types (multispectral images) and by applying it to other planetary surfaces. This methodology has a potential to become a powerful investigative tool with a wide range of applications. To facilitate its adoption by the research community the code that implements the SLT and its documentation will be put in the public domain. The results of this work will be disseminated through new courses, seminar talks, and collaborations with other institutes.