Translator Disclaimer
Paper
23 November 1982 Study Of Sensor Spectral Responses And Data Processing Algorithms And Architectures For Onboard Feature Identification
Author Affiliations +
Abstract
A computational model of the deterministic and stochastic processes involved in remote sensing is used to study spectral feature identification techniques for real-time onboard processing of data acquired with advanced Earth-resources sensors. Preliminary results indicate that: Narrow spectral responses are advantageous; signal normalization improves mean-square distance (MDS) classification accuracy but tends to degrade maximum-likelihood (MLH) classification accuracy; and MSD classification of normalized signals performs better than the computationally more complex MLH classification when imaging conditions change appreciably from those conditions during which reference data were acquired. The results also indicate that autonomous categorization of TM signals into vegetation, bare land, water, snow and clouds can be accomplished with adequate reliability for many applications over a reasonably wide range of imaging conditions. However, further analysis is required to develop computationally efficient boundary approximation algorithms for such categorization.
© (1982) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
F. O. Huck, R. E. Davis, C. L. Fales, and R. M. Aherron "Study Of Sensor Spectral Responses And Data Processing Algorithms And Architectures For Onboard Feature Identification", Proc. SPIE 0345, Advanced Multispectral Remote Sensing Technology and Applications, (23 November 1982); https://doi.org/10.1117/12.933770
PROCEEDINGS
24 PAGES


SHARE
Advertisement
Advertisement
Back to Top