A computer program has been written for use in maintaining the calibration, with respect to the positions of imaged objects, of a stereoscopic pair of cameras on each of the Mars Explorer Rovers Spirit and Opportunity. The program identifies and locates a known object in the images. The object in question is part of a Mössbauer spectrometer located at the tip of a robot arm, the kinematics of which are known.
In the program, the images are processed through a module that extracts edges, combines the edges into line segments, and then derives ellipse centroids from the line segments. The images are also processed by a feature-extraction algorithm that performs a wavelet analysis, then performs a pattern-recognition operation in the wavelet-coefficient space to determine matches to a texture feature measure derived from the horizontal, vertical, and diagonal coefficients. The centroids from the ellipse finder and the wavelet feature matcher are then fused to determine co-location. In the event that a match is found, the centroid (or centroids if multiple matches are present) is reported. If no match is found, the process reports the results of the analyses for further examination by human experts.
This program was written by Terrance Huntsberger, Yang Cheng, Robert Liebersbach, and Ashitey Trebi-Ollenu 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-43470.
This Brief includes a Technical Support Package (TSP).
Fusing Image Data for Calculating Position of an Object
(reference NPO-43470) is currently available for download from the TSP library.
Don't have an account?
Overview
The document discusses a software development project aimed at improving the localization of the Instrument Deployment Device (IDD) on the Mars Exploration Rover (MER). After the MER's landing in 2003, the front hazard avoidance cameras lost calibration, necessitating a method for re-calibration. Engineers utilized the rover's robotic arm to position the Mossbauer (MB) Spectrometer at known locations, allowing for the re-calibration of the cameras based on their kinematics.
To maintain camera calibration over time, a tool was developed to autonomously identify the center of the MB plate in images, independent of the camera model. This tool needed to account for image artifacts that could distort center computations, thus incorporating a sanity check to validate its findings. By comparing stereo images to known kinematics, a three-dimensional error vector could be derived, ensuring ongoing accuracy in camera positioning.
The software employs a multi-feature fusion approach, combining ellipse and wavelet features extracted from downlinked images of the contact ring on the Mossbauer imager. The process involves several steps: edge extraction, line segment combination, ellipse centroid derivation, and wavelet feature analysis. The fusion of these features allows for the automatic determination of the IDD end-effector position. If a match is found, the software reports the centroid; if not, it provides candidates for further analysis by operators.
The software has been tested on images from the MER Spirit and Opportunity rovers, demonstrating its effectiveness in determining the end-effector position. As the mission progresses, the aging of the IDD joints may lead to discrepancies between commanded and actual positions, making this software crucial for ongoing operations.
Overall, the document highlights the innovative techniques developed to enhance the operational capabilities of the MER, ensuring that the rover can continue its exploration of Mars effectively. The fusion of image data not only addresses the challenges posed by the Martian environment but also represents a significant advancement in autonomous robotic systems for space exploration.
