Table 4.

ANN algorithm results for single layer stacked image.

ANN results for single layer stacked image
Training 70							Testing 30
No of training		Fields_ training	Forest_ training	Shrubs_ training	Urban_ training	Overall accuracy	Fields_ testing	Forest_ testing	Shrubs_ testing	Urban_ testing	Overall accuracy
50	Overall accuracy = 91.23%, kappa coefficient = 0.864						Overall accuracy = 92.83%, kappa coefficient = 0.889
	Unclassified	0	0	0	0		0	0	0	0
	Fields	15,922	791	1019	304	93.11%	6891	292	137	189	93.41%
	Forest	207	31,004	1149	17	93.83%	106	13,280	290	12	93.88%
	Shrubs	284	1135	3337	10	60.62%	111	550	1631	10	79.25%
	Urban	688	114	0	9211	96.53%	269	24	0	3980	94.97%
100	Overall accuracy = 90.94%, kappa coefficient = 0.854						Overall accuracy = 92.54%, kappa coefficient = 0.884
	Unclassified	0	0	0	0		0	0	0	0
	Fields	15,567	382	436	787	91.03%	6748	106	49	416	91.47%
	Forest	457	31,397	1457	25	95.02%	209	13,518	318	17	95.56%
	Shrubs	856	1200	3612	19	65.61%	361	518	1691	14	82.17%
	Urban	221	65	0	8711	91.29%	59	4	0	3744	89.33%
200	Overall accuracy = 91.37%, kappa coefficient = 0.865						Overall accuracy = 93.77%, kappa coefficient = 0.902
	Unclassified	0	0	0	0		0	0	0	0
	Fields	16,385	778	1166	516	95.81%	7102	274	296	134	96.27%
	Forest	290	31,580	1733	25	95.57%	116	13,565	24	414	95.89%
	Shrubs	151	605	2606	5	47.34%	84	12	3866	0	92.25%
	Urban	275	81	0	8996	94.28%	75	295	5	1510	73.37%
300	Overall accuracy = 91.91%, kappa coefficient = 0.875						Overall accuracy = 93.48%, kappa coefficient = 0.899
	Unclassified	0	0	0	0		0	0	0	0
	Fields	16,291	574	800	596	95.26%	7067	187	329	71	95.8%
	Forest	153	30,836	819	16	93.32%	86	13,244	15	170	93.62%
	Shrubs	418	1555	3886	24	70.59%	67	13	3832	0	91.43%
	Urban	239	80	0	8906	93.33%	157	702	15	1817	88.29%