The present invention relates to a method for detecting a target material which detects a distribution of a target material which is targeted from a hyperspectral image for a specific region. According to the present invention, the method for detecting a target material using a hyperspectral image detects a distribution of a target material which is targeted from a hyperspectral image for a specific region to resolve problems of conventional methods for detecting a target material which determine a target material in pixel units to incorrectly detect a different material which is not the target material as the target material or often cannot detect a small target material. The method for detecting a target material using a hyperspectral image acquires a hyperspectral image for a region to be analyzed, automatically extracts an image endmember from the acquired hyperspectral image, then adds a target spectrum for the target material as a different endmember to perform spectral unmixing, and obtains an occupation percentage for the target material based on a spectral unmixing result to create a distribution map for the target material. The method for detecting a target material using a hyperspectral image based on sub-pixels can accurately detect only the target material which is targeted with almost no incorrectly detected pixels.;COPYRIGHT KIPO 2016
展开▼