Hybrid model to optimize object-based land cover classification by meta-heuristic algorithm: an example for supporting urban management in Ha Noi, Viet Nam

Figures & data

Figure 1. Study area.

Table 1. SPOT 7 specification.

	Bands	Spectral range	Spatial resolution
Original image	Blue	0.455 µm – 0.525 µm	6.0 m
	Green	0.530 µm – 0.590 µm	6.0 m
	Red	0.625 µm – 0.695 µm	6.0 m
	Near-Infrared	0.760 µm – 0.890 µm	6.0 m
	Panchromatic	0.450 µm – 0.745 µm	1.5 m
Fusion	(B,G,R,NIR)		1.5 m

Figure 2. Image segmentation at Scale: 30, Shape: 0.75, and Compactness: 0.5.

Table 2. Predictor variables extracted from image segmentation.

ID	Object features	Number of bands	Description
1.	Min pixel value	6	Min pixel values for each of 6 bands
2.	Max pixel value	6	Max pixel values for each of 6 bands
3.	Mean pixel value	6	Mean pixel values for each of 6 bands
4.	Standard deviation	6	Standard deviation of pixel value
5.	Min_PP	6	Mean pure pixel
6.	Max_PP	6	Max pure pixel
7.	Mean_PP	6	Mean pure pixel
8.	Standard deviation _PP	6	Standard deviation of pure pixel
9.	Circular	1	Circular
10.	Compactness	1	Compactness
11.	Elongation	1	Elongation
12.	Rectangular	1	Rectangular
Total predictor variables		52

Figure 3. Neural network topology.

Table 3. Parameters of GOA, GWO, and BBO.

Parameters	GOA	BBO	GWO
Number of agents (population size)	100 (Number of grasshoppers)	100 (Number of habitats)	100 (Number of Wolfs)
Variable dimensions	560: Positions of grasshoppers in 560-dimensional space	560 (Suitability Index Variables – SIV)	560: Positions of Wolfs in 560-dimensional space
Fitness function	RMSE by Equation (3)	RMSE by Equation (3)	RMSE by Equation (3)
Optimal value	The location of the best agent (grasshopper)	Highest Habitat Suitability Index (HIS)	The location of Alpha Wolf
Specific parameters	cMax = 1; cMin = 0.00004	Emigration rate (ER) is defined by population size Immigration rate (IM) = 1 – ER Mutation rate = 0.1 Keep rate = 0.3	Parameters are initialized are updated during the optimization course.
Maximum iteration	500	500	500
Variable lower and upper bounds	[−1,1]	[−1,1]	[−1,1]

Table

		Predicted value
		Yes	No
Actual value	Yes	True Positive (TP)	False Negative (FN)
	No	False Positive (FP)	True Negative (TN)

Figure 4. Prototype model.

Figure 5. Neural network optimized GOA after 500 iteration.

Table 4. Accuracy assessment for 4000 training objects (best values in shaded cells).

Classifier/Base classifier	Optimization algorithm	RMSE	MAE	Kappa	Overall Accuracy
MNN	GWO	0.2477	0.0902	0.7685	83.2381
MNN	BBO	0.2335	0.1129	0.7843	84.4762
MNN	GOA	0.2225	0.1086	0.8284	87.1905
DT		0.2345	0.1244	0.7848	84.4771
SVM		0.2564	0.1221	0.7503	82.0012

Figure 6. ROCs and AUC values for 4000 training objects.

Figure 7. ROCs and AUC values for 2000 training objects.

Figure 8. ROCs and AUC values for 400 training objects.

Table 5. Accuracy assessment for 2000 training objects (best values in shaded cells).

Classifier/Base classifier	Optimization algorithm	RMSE	MAE	Kappa	Overall Accuracy
MNN	GWO	0.2337	0.0846	0.798	0.853333
MNN	BBO	0.2241	0.112	0.799	0.854667
MNN	GOA	0.2217	0.0866	0.8136	0.869300
DT		0.2375	0.1309	0.7759	0.838667
SVM		0.258	0.1238	0.7348	0.809333

Table 6. Accuracy assessment for 400 training objects (best values in shaded cells).

Classifier/Base classifier	Optimization algorithm	RMSE	MAE	Kappa	Overall Accuracy
MNN	GWO	0.2826	0.1033	0.7242	0.802
MNN	BBO	0.2447	0.1043	0.777	0.840
MNN	GOA	0.2373	0.1029	0.7836	0.844
DT		0.2414	0.1299	0.7715	0.836
SVM		0.267	0.1316	0.7353	0.806

Table 7. Wilcoxon signed-rank test.

Pairwise comparison	p-value (<0.05)	Significance
GMNN vs. GWO-MNN	0.0004883	Yes
GMNN vs. BBO-MNN	0.003216	Yes
GMNN vs SVM	0.0009766	Yes
GMNN vs. DT	0.007602	Yes

Figure 9. Subsets of classified maps by GMNN with different training and validation sizes. From left to right: 4000 objects, 2000 objects, 400 objects.

Table 8. Sample of probability assignments.

Predicted classes	Prob of being Vegetation	Prob of being Water	Prob of being Bare soil	Prob of being Impervious
water	0	0.973	0.026	0
bare soil	0.222	0.098	0.655	0.024
vegetation	0.955	0	0	0.045
vegetation	0.967	0	0	0.033
vegetation	0.986	0.001	0	0.013
impervious	0.002	0.095	0	0.903
water	0.036	0.832	0.014	0.118
water	0	0.672	0.328	0