Karst vegetation coverage detection using UAV multispectral vegetation indices and machine learning algorithm

Table 5 Mutual validation parameters of optimal models at different flight altitudes

GBM model flight altitude (m)	Data flight altitude (m)	Classification	R (%)	F1 (%)	OA (%)
100	All	gv	74.95	85.03	80.41
		ro	95.53	59.22
		so	70.63	78.87
		wd	91.99	90.60
	200	gv	72.58	83.71	83.19
		ro	99.54	82.92
		so	82.52	82.52
		wd	95.56	82.71
	400	gv	90.35	93.69	84.96
		ro	69.23	62.61
		so	27.40	35.85
		wd	93.89	76.99
200	All	gv	98.76	93.53	89.03
		ro	88.76	87.37
		so	52.43	66.74
		wd	75.86	82.97
	100	gv	99.36	96.31	94.78
		ro	95.74	97.01
		so	86.01	92.27
		wd	90.48	91.39
	400	gv	97.83	96.72	88.82
		ro	63.08	67.91
		so	25.48	34.08
		wd	88.55	80.48
400	All	gv	87.77	92.53	80.63
		ro	91.15	74.71
		so	64.07	47.77
		wd	52.21	68.31
	100	gv	98.44	98.31	80.21
		ro	97.01	69.61
		so	39.91	49.01
		wd	52.84	69.14
	200	gv	78.58	87.04	66.15
		ro	87.96	53.45
		so	30.21	25.78
		wd	38.02	55.09
All	100	gv	97.42	97.92	95.66
		ro	96.72	94.10
		so	85.27	91.17
		wd	97.26	94.90
	200	gv	84.30	91.41	89.79
		ro	99.08	85.89
		so	81.15	78.06
		wd	97.39	92.82
	400	gv	93.27	95.58	88.31
		ro	79.69	74.59
		so	36.89	47.35
		wd	92.29	78.24

All represents the total including 100 m, 200 m and 400 m data
Bold represents the highest overall accuracy when the optimal models at different flight altitudes were mutually validated

ISSN: 1746-4811