| 摘要 |
A method for processing spectral image data, comprising: identifying two endmembers represented in the spectral image data, estimating abundance fraction values of the identified endmembers, reconstructing the spectral image based on the estimated abundance fraction values, determining a difference image between the initial spectral image and the reconstructed image, applying principal component analysis to the difference image, wherein applying PCA comprises performing one or more iterations of a PCA algorithm, storing a linear function and associated vectors generated by each iteration, storing the abundance image and endmembers, compressing the linear function and associated vectors, abundance image and endmembers. |
