AI Remote Sensing: Development of Fuzzy Model for Enhancing Land Cover Classification

Getting accurate classification is a challenging task. This study introduces the concept of AI remote sensing, specifically the development of a fuzzy fusion model using Dempster-Shafer theory to enhance land cover classification in the Damietta governorate. The study demonstrates the effectiveness of the fuzzy fusion model in improving the accuracy of land cover classification, particularly for the building class. Firstly, three benchmark supervised classifiers (Normal Bayes Classifier, Support Vector Machines, and KNN) were selected and implemented. Next, according to the output of the classifiers, the classification results were fused by the Dempster-Shafer fuzzy model. Finally, the classifiers’ performances were evaluated using the confusion matrix. Results indicated that the accuracy of the building class was improved using fuzzy methods.