Improvement in the Accuracy of the Postclassification of Land Use and Land Cover Using Landsat 8 Data Based on the Majority of Segment-Based Filtering Approach

Table 2

Results of the accuracy assessment for the LULC classification based on the ML classifier without the MaSegFil approach.

LULC class	C_0	C_1	C_2	C_3	C_4	C_5	C_6	C_7	C_8	C_9	C_10	Total	User accuracy	Kappa
C_0	105	22	2	0	0	2	0	0	0	0	0	131	0.802	0.000
C_1	32	199	27	0	0	10	0	0	1	0	0	269	0.740	0.000
C_2	24	19	64	0	0	4	25	21	19	0	0	176	0.364	0.000
C_3	11	0	1	5	0	0	37	1	24	0	0	79	0.063	0.000
C_4	3	5	1	0	3	0	1	11	0	2	0	26	0.115	0.000
C_5	27	49	22	0	0	33	2	1	6	0	0	140	0.236	0.000
C_6	12	0	2	0	0	0	551	6	12	0	0	583	0.945	0.000
C_7	1	0	2	0	0	0	31	357	6	1	0	398	0.897	0.000
C_8	13	0	12	0	0	1	25	8	145	0	0	204	0.711	0.000
C_9	0	0	0	0	0	0	0	4	0	59	5	68	0.868	0.000
C_10	0	0	0	0	0	0	0	0	0	8	45	53	0.849	0.000
Total	228	294	133	5	3	50	672	409	213	70	50	2127	0.000	0.000
Procedure accuracy	0.461	0.677	0.481	1.000	1.000	0.660	0.820	0.873	0.681	0.843	0.900	0.000	0.736	0.000
Kappa	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.684

C_0: primary dryland forest; C_1: secondary dryland forest; C_2: fields; C_3: open field; C_4: wetland forest; C_5: plantation; C_6: settlement; C_7: rice fields; C_8: shrubs; C_9: fishpond; C_10: water body.