Absorption Band Spectrum Features Extraction for Minerals Recognition Based on Local Spectral Continuum Removal

Hyperspectral mineral identification methods have a wide application in field. In this paper, we propose a new pipeline of mineral identification by using absorption band spectrum features. In the pre-processing, a local spectral continuum removal algorithm is used to normalise the corresponding spectral data of the absorption band of different minerals. After that, the polynomial fitting is applied to remove the spectral outliers. Then, by using the preprocessing spectral data, the absorption band spectrum features are extracted to establish a rules based interring system for minerals recognition. Experimental results on hyperspectral images demonstrate that the proposed method has good performance in minerals recognition.