Enhancing Satellite Image Retrieval with RNN Algorithm and DNN -Based Bat Optimization Classifier

Remote sensing has been predominantly used in assessing environmental changes, disaster management, and studies of urban structure, among others, and as the quality of satellite images increases, methods for efficient storage and retrieval become important. This research finds itself right at addressing the challenge of satellite image retrieval through proposing an efficient search method that comprises RNN based algorithm and classifier, as well as DNN based BAT optimization. The methodology builds on RNN, notably LSTM to capture temporal dependencies and to identify relevant features of satellite image sequences. This in turn helps to improve the retrieval efficiency of an article and a DNN-based BAT optimization classifier is applied to fine-tune the weights and bias terms of the network. Based on an extensive broad satellite image dataset used in the experimental assessment, the effectiveness enhancement was confirmed. The proposed method gives an F1 score of 92.4% which shows that the proposed method yields a superior result over the compared method such as RNN-, CNN BAT and NN classifiers. The findings establish that applying RNN algorithms with BAT optimization in constructing satellite image search system is accurate and efficient to establish a sophisticated geographical data analysis, which holds potential for future applications.