Remote Sensing Image Classification Based on Confidence Score of Ensemble Machine Learning Classifiers

A novel algorithm for classifying remote sensing images is presented in this paper. The proposed methodology uses confidence score with ensemble of classifiers to categorize the satellite images. Images from remote sensing satellite dataset that includes image categories of Cloudy, Dessert region, Water cover and Green area are initially preprocessed for enhancing the quality. Four classifiers namely Decision tree, Random Forest, Gaussian Naïve Bayes and Support vector classifiers are trained individually with the preprocessed images. Based on the training accuracy and loss, confidence scores are computed for each classifier. The final classification label is decided with the aid of confidence score of ensembles of classifies. The proposed methodology is validated both quantitatively and qualitatively. Overall accuracy of approximately 92% is achieved with minimal loss for each category of image. Comparison of the proposed method with various state of the art algorithms proves that the proposed algorithm with ensemble of classifiers performs better in classifying the images.