A learning-based target decomposition method using Kernel KSVD for polarimetric SAR image classification

EURASIP Journal on Advances in Signal Processing

Table 2 Classification accuracy of Kernel KSVD and other decomposition methods based on different polarimetric matrixes for rice data (the bolded value represents the maximum accuracy among Kernel KSVD and the responding comparison methods for each ground object and each polarimetric matrix)

Matrix	Feature	rice1	rice2	rice3	rice4	rice5	Accuracy
	Kernel KSVD _[S]	98.64	94.08	95.05	95.65	90.76	96.49
	Coherency	89.95	53.76	65.95	66.05	71.73	75.55
Matrix [S]	Krogager	96.73	91.31	89.76	89.64	85.34	92.91
	Pauli	98.35	92.02	92.09	93.07	90.18	95.02
	Kernel KSVD_[C]	98.67	90.09	91.01	92.55	86.91	94.48
	OEC	96.27	93.42	89.62	87.11	86.58	92.53
Matrix [C]	FourComponent	97.87	94.37	93.76	93.73	90.12	95.55
	Freeman	97.90	92.37	92.39	93.22	90.66	94.97
	Kernel KSVD_[T]	98.90	92.47	92.47	94.19	87.01	95.78
	Cloude	94.16	84.43	84.29	84.63	69.29	88.24
Matrix [T]	Huynen	97.26	90.63	90.93	91.82	88.89	93.85
	Holm	96.97	92.14	90.88	90.83	82.34	93.44
	Barnes	98.37	94.06	94.06	95.40	90.56	96.13