FIGURE 2. Texture images.
Next, the generated texture images (Fig. 2) were classified
respectively by the proposed new fuzzy clustering remote sensing
classification method with neighborhood distance constraint and the
results are shown in Fig. 3. Within the classification results, the
classes‘ number c = 4, fuzzy factor m = 3, suspensive condition e =
0.0002, neighborhood weight a = 0.7, and the number of
neighborhood pixels NR = 8.
As we can observe from Fig. 3, the fuzzy cluster classification
(Fig. 3a) is better than the other three texture images (Fig. 3b–d).
From Fig. 3(a), the feature points of texture image are relatively and
uniformly classified. According to the analysis, the texture image
was smoothed and the number of the cluster centers was decreased