Improving the Accuracy of Land Use/Cover Maps using an Optimization Technique

Mapping of Land use/cover is important for many activities of planning and management, especially in arid areas. Nowadays, satellite imagery and remote sensing techniques, which provide timely and high capability data, are widely used in producing this kind of mapping. The main objective of this study is to produce an actual land use map using advanced pixel-based (MLP, SVM, and SOM) approaches. The most important challenge, in this case, is to determine the optimum structure of classification methods. The Taguchi method is used to optimize the structure of MLP, SVM, and SOM methods. Results show that the Taguchi method can be effectively used to cope with this problem. It significantly reduces the number of classification tests. We also showed that there are significant differences between the results of the SVM method with those of the MLP and SOM methods (χ2 more than 100) and that SVM model is more efficient than other methods. The accurate map produced by the optimized SVM approach (Overall accuracy of 0.93) showed that this method has a better performance.