A Method Combining CNN and ELM for Feature Extraction and Classification of SAR Image

Table 2

Recognition and contrast results (%) of different algorithms for SAR images.

Accuracy rate	SRC; [2]	DCN; [4]	CGLF; [5]	PCA; [1]	CSPC; [3]	CNSVM; [25]	CNN-; SVM; [7]	CNNELM
2S1	85.36	98.18	95.62	89.95	90.65	94.53	100.00	100.00
BMP_2	84.48	98.98	98.46	94.54	92.5	95.97	100.00	100.00
BRDM_2	90.11	99.27	95.26	92.54	93.34	85.04	100.00	100.00
BTR_60	96.53	96.41	100.00	95.68	98.95	97.46	99.99	100.00
BTR_70	98.33	99.49	99.49	85.66	99.79	91.33	97.66	100.00
D_7	98.32	99.27	100.00	90.85	98.12	98.18	100.00	100.00
T_62	90.54	99.64	98.17	89.99	90.52	91.21	99.50	100.00
T_72	94.59	100.00	100.00	96.79	91.38	98.18	99.50	100.00
ZIL_13_1	96.35	99.64	99.49	94.68	94.65	93.80	99.00	100.00
ZUS_23_4	92.3	99.64	99.64	93.65	94.62	98.18	100.00	100.00
Average	92.70	99.13	98.56	92.43	94.48	94.38	99.57	100.00