Title :
Accuracy Analysis on Sub-pixel Mapping Using Normalized Fourier Descriptor
Author_Institution :
Sch. of Hydropower & Inf. Eng., Huazhong Univ. of Sci. & Technol., Wuhan, China
Abstract :
Sub-pixel mapping is a vital process on the unmixing and locating of remotely sensed mixed pixels. Each mixed pixel decomposes to S×S end members, in which S delegates spatial scale factor. How to choice the S value directly determines positioning accuracy and reliability. This paper focuses to the accuracy analysis on sub-pixel mapping by applying normalized Fourier descriptor to describe and assess its precision using different S values. Experiments show that when S≤3 the sub-pixel position maps are reliable, when S=4, 5 they can be applied in some low accuracy application fields, when S≥6 they are unreliable.
Keywords :
Fourier transforms; image processing; accuracy analysis; normalized Fourier descriptor; remotely sensed mixed pixels; sub-pixel mapping; Accuracy; Classification algorithms; Educational institutions; MATLAB; Pattern recognition; Remote sensing; Shape; accuracy analysis; mixed pixel; normalized Fourier descriptor; sub-pixel mapping; sub-pixel position map;
Conference_Titel :
Computer Sciences and Applications (CSA), 2013 International Conference on
Conference_Location :
Wuhan
DOI :
10.1109/CSA.2013.120