Date of Original Version
Abstract or Description
Future planetary rovers will travel autonomously over the visible horizon where data collection sequences cannot be scripted in advance. These rovers would benefit from an ability to autonomously identify science targets and deploy sensors. We present an automatic method for measuring rocks using a VIS/NIR reflectance spectrometer and solar illumination. Our approach leverages an automatic image-based rock detection algorithm and SIFT keypoints  that establish feature correspondence across images. This permits a rover to collect multiple spectral target measurements in a single command cycle with no direct human intervention. We present field tests of the system at Amboy Crater in the Mojave Desert. Here rock detection with visual servoing significantly improves the diversity and purity of collected spectra.
Proceedings of the 9th International Symposium on Artificial Intelligence, Robotics and Automation in Space (iSAIRAS).