Rockster is an algorithm that automatically identifies the locations and boundaries of rocks imaged by the rover hazard cameras (hazcams), navigation cameras (navcams), or panoramic cameras (pancams). The software uses edge detection and edge regrouping to identify closed contours that separate the rocks from the background (see figure). The algorithm has applications both in ground-based data analysis, for example, to examine large quantities of images returned by the Mars Exploration Rovers, and in onboard (on-rover) opportunistic science applications such as construction of rock maps during traverse, identification of unusual or otherwise high-value science targets that warrant additional investigation, and detection of certain types of geologic contact zones.
The algorithm is particularly efficient at quickly detecting small- to mediumsized rocks with sufficient contrast (positive or negative) relative to the background. Full quantitative performance comparisons are not yet available; however, preliminary tests show that Rockster appears to detect a significantly larger fraction of rocks present in a scene (higher recall) than previous rock detection schemes, while maintaining a high precision rate (objects identified as rocks, truly are rocks).
Rockster has been integrated successfully into a number of recent, high-level demonstrations, including the SOOPS (Science Operations on Planetary Surfaces) demo, which used a rock exploration scenario to let scientists gain hands-on experience with an autonomous science capability in a simulated environment, and live exercises of the OASIS (Onboard Autonomous Science Investigation System)/CLARAty (Coupled Layer Architecture for Robotic Autonomy) software which were carried out in real-time in the JPL Mars Yard onboard the FIDO Rover (a close relative of the twin Mars Exploration Rovers).
This program was written by Michael Burl of Caltech for NASA’s Jet Propulsion Laboratory. This software is available for commercial licensing. Please contact Karina Edmonds of the California Institute of Technology at (626) 395-2322. Refer to NPO-44417.