engleski

Method for evaluation of laboratory craters using crater detection algorithm for digital topography data

In our previous work the following has been done: (1) the crater detection algorithm (CDA) based on digital elevation model (DEM) has been developed and the GT-115225 catalog has been assembled [GRS, 48 (5), in press, doi:10.1109/TGRS.2009.2037750] ; and (2) the results of comparison between explosion-induced laboratory craters in stone powder surfaces and GT-115225 have been presented using depth/diameter measurements [41stLPSC, Abstract \#1428]. The next step achievable using the available technology is to create 3D scans of such laboratory craters, in order to compare different properties with simple Martian craters. In this work, we propose a formal method for evaluation of laboratory craters, in order to provide objective, measurable and reproducible estimation of the level of achieved similarity between these laboratory and real impact craters. In the first step, the section of MOLA data for Mars (or SELENE LALT for Moon) is replaced with one or several 3D-scans of laboratory craters. Once embedment was done, the CDA can be used to find out whether this laboratory crater is similar enough to real craters, as to be recognized as a crater by the CDA. The CDA evaluation using ROC' curve represents how true detection rate (TDR=TP/(TP+FN)=TP/GT) depends on the false detection rate (FDR=FP/(TP+FP)). Using this curve, it is now possible to define the measure of similarity between laboratory and real impact craters, as TDR or FDR value, or as a distance from the bottom-right origin of the ROC’ curve. With such an approach, the reproducible (formally described) method for evaluation of laboratory craters is provided.

evaluation of laboratory craters; crater detection algorithm; digital topography data

nije evidentirano