Fig. 1. Spectra of spectral matching operator based on position vector statistics. (a) Original spectral curves; (b) results of SAM; (c) results of SID; (d) results of SCM

Fig. 2. Flowchart of voting statistics

Fig. 3. Flowchart of operators

Fig. 4. Pixel spectral curve and fitting curve of eigenmatrix

Fig. 5. Visualization results of operator. (a) Pixels of the same class; (b) pixels of different classes

Fig. 6. ROC curves

Fig. 7. Target location diagram. (a) Data 1; (b) data 2

Fig. 8. Recognition effects of single operator in data 1. (a) SAM; (b) SID; (c) SCM; (d) PVS

Fig. 9. Recognition effects of single operator in data 2. (a) SAM; (b) SID; (c) SCM; (d) PVS

Fig. 10. ROC curves. (a) Data 1; (b) data 2

Fig. 11. Two types of pixels with similar spectral curves

Fig. 12. Recognition effects of data 1. (a) CEM; (b) OSP; (c) SAM,SID,SCM fusion; (d) SAM,SID,SCM,PVS fusion

Fig. 13. Recognition effects of data 2 . (a) WCM-CEM; (b) WCM-OSP; (c) SAM,SID,SCM fusion; (d) SAM,SID,SCM,PVS fusion

Fig. 14. ROC curves. (a) Data 1; (b)data 2