When classifying the overall image based on the Tent-PSO-SVM model, another two comparative experiments were also designed: i.e., SVM classification versus PSO-SVM classification, and the final classification results obtained from the three models are shown in Plate 4.1 to 4.3, respectively.
The category information marked by the ellipse in the lower right corner of Plate 4.2 and Plate 4.3 is obviously finer than that in the same position of Plate 4.1, which also indicates that selecting the optimal combination of bands for SVM classification of hyperspectral imagery is necessary in improving the classification accuracy. The left-hand airfield located in the middle of Plate 4.3, marked by the ellipse, is still partially recognized, but the airfield in the same location of Plate 4.2 is not recognized, which can also reflect the advantage of the improved Tent mapping acting on the PSO algorithm for SVM classification of hyperspectral images.