Abstract
An approach for identification of objects without specified shape using multispectral satellite data is developed. The approach is based on the orthogonalization procedure in extended multidimensional spectral space in which filters that are orthogonal to the hypotheses of desired and similar objects are calculated and, then, scalarly multiplied by the spectra under study. It is shown that normalized rather than original spectra must be used, which leads to a significant decrease in variability of spectra that results in worse recognition of objects and specific calibration must be employed for suppression of atmospheric distortions. It is shown that such a method provides significantly higher recognizability in comparison with the least-squares method that is used in most recognition techniques.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
