Multi-Sensor Terrain Fusion

TSC will develop rigorous mathematical techniques to combine digital terrain elevation data from diverse multi-sensor sources to automatically generate a composite Digital Elevation Map (DEM). Sensor sources to be included for Phase I algorithmdevelopment and evaluation include existing DEMs, stereo photogrammetry, LIDAR, IFSAR, and stereo SAR. Quality metrics and error models will be developed for input to advanced regridding techniques which are now being used for IFSAR ground processing.These techniques will be modified to fuse diverse sources of terrain height measurements into an optimal composite DEM. TSC will also investigate co-kriging and novel belief propagation techniques, and investigate methods to detect and remove systematicerrors, including tilts, shifts and biases which plague many sensor data sources. TSC will then design a preliminary full-scale software architecture that is compatible with both Navy and commercial operations.During Phase I, TSC will utilize GFE or in-house multi-sensor terrain data including aerial photos, X/P-band IFSAR and UHF/VHF stereo SAR to demonstrate the feasibility of these techniques. In Phase II, TSC will extend the most promising techniques toprocess data from additional sensors, of interest to the Navy, and will develop software algorithms that can be integrated with existing Navy and NIMA DEM generation tools.