A mosaic is a commonly used tool for representing the enormous amount of data generated from video sequences. In contrast to the usual application of mosaics as a user interface, the forestry environmental monitoring domain requires accurate geo-corrected mosaics tied to real-world coordinates. The standard techniques of generating seamless mosaics using only image data in a frame-by-frame image registration process has serious deficiencies over extended periods of time due to error accumulation, even when errors between two successive frames are very small. Our mosaics require both seamless registration of optical data, and the use of flight sensor data (geo-data) to provide a globally correct track of the motion to keep errors from propagating. A parallel- perspective projection mosaic representation (P3 mosaic) is proposed to represent the geo-corrected mosaic. Our complete mosaic method includes local registration, track generation, matching refinement, and a two-track-based mosaic composition.
Free mosaic: Seamless mosaic, but not geo-corrected; Notice the
accumulating error of scale
Edward M. Riseman, ProfessorOthers:
Allen R. Hanson, Professor
Howard Schultz, Senior Research Scientist
Frank Stolle, Ph.D. student
Chris Holmes, system programmer
Dana M. Slaymaker, Department of Wildlife and Forestry Conservation, UMass
Chris Hayward, graduate student, Department of Wildlife and Forestry Conservation, UMass
National Science Foundation Project (Grant Number EIA- 9726401), Automatic Interpretation of High-Altitude Image Data for Eco-System Modeling, $1,800,000, 02/01/98 – 01/31/01, PI (Riseman), Co-PIs (Hanson, Slaymaker)