SM-EEAs are always desirable for endmember extraction. Unfortunately, a genuine SM-EEA is generally impractical because of its very high computational cost resulting from an exhaustive search, specifically when the number of endmembers, p, is large and data volume is huge. In addition, an SM EEA also suffers from several drawbacks: (1) requirement of precise knowledge about p, which is practically unknown; (2) assumption of endmembers present in the data, which is generally not true in many real applications; (3) necessity of dimensionality reduction (DR) due to enormous data volume, in which case selecting an effective DR transform is crucial; and (4) inconsistent results caused by the use of randomly generated initial endmembers. While some drawbacks such as p that can be estimated by the virtual dimensionality in Chapter 5, some other drawbacks such as computational complexity inheriting from the algorithm design. Table 7.1 summarizes design criteria of SM-EEAs discussed in this chapter and their drawbacks and disadvantages.
SM-EEAs | Design criteria | Drawbacks/disadvantages |
MATLAB-PPI | Orthogonal projection | 1. Determination of the number of skewers, K
2. Determination of a threshold for PPI counts, t 3. Determination of the number of dimensions, q, for DR 4. Inconsistent results due to the use of random conditions 5. Many falsely extracted endmembers |
SM N-FINDR, IN-FINDR, ASM N-FINDR | Maximal simplex volume | 1. Precise knowledge about p
2. Determination of the number of dimensions, q, for DR 3. High computational complexity 4. Inconsistent results due to the use of random conditions 5. Assumption on presence of pure signatures |
MVT | Minimal simplex volume | |
CCA | Convex cone | 1. Precise knowledge about p
2. Determination of the number of dimensions, q, for DR 3. High computational complexity 4. Assumption on presence of pure signatures |
SPCA-EEA | Statistical spectral correlation | 1. Precise knowledge about p
2. Inconsistent results due to the use of random conditions 3. Less effective |
FCLS-EEA | Linear spectral unmixing | 1. Precise knowledge about p
2. High computational complexity 3. Ill-rank of endmember signature matrix 4. Inconsistent results due to the use of random conditions |
AMEE | Morphology | 1. Precise knowledge about p
2. All endmembers of the same type are extracted before endmembers of another type 3. Inconsistent results due to the use of random conditions 4. A process is needed to discriminate extracted endmembers |
3.133.141.219