A Two-Stage Optimization Strategy for Fuzzy Object-Based Analysis Using Airborne LiDAR and High-Resolution Orthophotos for Urban Road Extraction

Table 4

and orthogonal arrays and the POF used to optimize the shape and compactness parameters.

Run	Level 1			Level 2			Level 3
	MI	Variance	POF	MI	Variance	POF	MI	Variance	POF
(1)	0.396	0.930	1.327	0.599	0.964	1.563	1.464	1.011	2.474
(2)	0.233	1.000	1.233	0.795	0.971	1.765	0.373	0.885	1.258
(3)	0.438	0.865	1.303	0.647	0.963	1.610	0.320	0.000	0.320
(4)	0.141	0.411	0.552	0.523	0.000	0.523	1.000	0.947	1.947
(5)	0.371	0.962	1.333	0.859	0.956	1.815	0.000	0.904	0.904
(6)	0.000	0.967	0.967	0.704	0.986	1.690	0.523	0.616	1.138
(7)	1.000	0.697	1.697	0.692	0.901	1.593	0.326	0.945	1.271
(8)	0.174	0.000	0.174	0.698	0.049	0.746	0.969	1.000	1.969
(9)	0.013	0.954	0.967	0.668	1.000	1.668	0.404	0.332	0.736
(10)	0.565	0.801	1.366	0.113	0.958	1.072
(11)	0.578	0.758	1.336	0.783	0.998	1.782
(12)	0.867	0.274	1.141	0.738	0.415	1.153
(13)	0.329	0.952	1.280	0.554	0.978	1.532
(14)	0.035	0.891	0.927	0.000	0.979	0.979
(15)	0.307	0.798	1.105	0.434	0.948	1.381
(16)	0.060	0.207	0.268	1.000	0.480	1.480