CRATER DETECTION TECHNIQUES ON DEMs FOR AUTOMATIC GENERATION OF LUNAR SURFACE DATABASE IN OPTICAL TERRAIN ABSOLUTE NAVIGATION
- Paper number
IAC-11,A3,2.P,12,x10488
- Author
Dr. Marco Mammarella, GMV, Spain
- Coauthor
Mr. Marcos Avilés, GMV, Spain
- Coauthor
Mrs. Ana Maria Sanchez Montero, GMV, Spain
- Coauthor
Mr. Pablo Colmenarejo, GMV, Spain
- Year
2011
- Abstract
Creating a reliable landmarks database is a crucial step in a Optical Terrain Absolute Navigation (OTAN) system, taking into account the entire spacecraft position estimation is relative to the on-board stored database. However, after decades of research and manual efforts, only tens of thousands of the millions of craters on the Moon have been catalogued, mainly, craters with diameters ≥ 2 km. Automated techniques for crater detection and cataloguing are therefore necessary to take advantage of the vast quantities of remotely sensed data now available, especially now that precise 3D information is becoming available. The main contribution of this effort is an algorithm that not only estimates the crater center and radius to the steepest border, but that it also provides a pixel-based accurate representation of the crater rim. The core of the described technique is to identify, for each crater, a point or region inside the crater and another one outside. In order to estimate the inside and outside regions, the Circular Hough Transform (CHT) is used with the purpose of identifying almost circular patterns on a binary image obtained from thresholding the gradient. The approximate center extracted using the CHT is used as the inside region and the approximate radius (multiplied by a confidence factor) is used as the outside region. The flooding mechanism, which is analog to a watershed transformation initialized using two markers, is used in order to find the craters rims. The algorithm has been totally implemented and the results are very promising, especially when dealing with older craters whose shape cannot be fitted to a circle or ellipse (which is the actual limitation of existent techniques). The output of the algorithm is the precise contour of the craters that can be approximated by circles, ellipses or a higher order functions depending of the final application. In the present paper the algorithm is tested accuracy and number of the identified craters and then the statistic distribution extracted is compared with the applied. Finally, the algorithm is tested with real DEMs representing specific zones and the results are compared with the known position of significant craters.
- Abstract document
- Manuscript document
(absent)