Feature Selection for Object-Based Classification of High-Resolution Remote Sensing Images Based on the Combination of a Genetic Algorithm and Tabu Search

Table 8

Classification accuracies for the QuickBird image.

Class	GATS		GA		TS		ReliefF
	PA (%)	UA (%)	PA (%)	UA (%)	PA (%)	UA (%)	PA (%)	UA (%)
Buildings	86.67	92.86	80.00	72.73	76.67	69.70	83.33	62.50
Vegetation 1	91.72	88.08	86.21	85.62	66.90	76.98	68.97	76.92
Vegetation 2	86.67	90.91	83.33	85.03	80.00	72.73	79.33	73.46
Water	86.67	86.67	73.33	84.62	66.67	90.91	66.67	90.91
Bare land	84.62	78.42	80.00	66.67	73.33	68.75	66.67	50.00
Sec. bare land	86.67	86.67	86.67	81.25	86.67	76.47	66.67	76.92
Roads	90.00	85.71	80.00	84.21	80.00	80.00	55.00	84.62
Shadows	80.00	87.50	60.00	75.00	90.00	75.00	80.00	72.73