Content based image retrieval using hybrid feature extraction and HWBMMBO feature selection method

Content Based Satellite Image Retrieval system is used for images processing, weather forecasting, climate monitoring, disaster management, forest fire detection etc. In this research forest, the fire retrieval approach is focused on protecting the forests from fire incidents which can generate an impact on natural resources and living organisms. This research proposed forest fire retrieval approach using proposed hybrid feature extraction technique and Hierarchy Weighted-Brownian Motion Monarch Butterfly Optimization Algorithm based feature selection approach. For application basis, both the proposed feature extraction and feature selection approach are implemented in forest fire retrieval system. The performances of the forest fire retrieval approach are analyzed in terms of precision, recall and accuracy. The efficiency of the proposed approach is evaluated by varying the features such as ground temperature, Top of Atmosphere, Land Surface Temperature, water vapour and intensity. The recall of the proposed fire retrieval approach is increased by 1.68%, 0.6%, 0.62%, 0.54% and precision by 0.82%, 0.41%, 0.75%, and 0.37% when compared with active fire detection, Convolutional Neural Network and hybrid intelligent algorithm respectively. The accuracy of the proposed fire retrieval approach is 98.91% better than the existing approaches.