An efficient content-based satellite image retrieval system for big data utilizing threshold based checking method

In the image processing as well as analysis field, Content-Based Satellites Image Retrieval (CBSIR) is a vitalissue. Though there are numerousprevailing Image Retrieval (IR) methods, they still need improvement in the retrieval accuracy along with computational intricacy.Thus, this paper proposed an efficient CBSIR system utilizingWeighted Brownian Motion-based Monarch Butterfly Optimizations(WBMMBO). Initially, the Satellite Images (SI)is taken as the input. On account of the explosive augmentation of SI, the dataset is larger, which in turn increases the requisite for attaining the best retrieval system. Next, the Adjusted Intensity-based Variant of Adaptive histograms equalization (AIVA) enhances the images’ contrast. After that, the LPDF, DCD, BoVW, SF, along with BRIEF features are extracted. Then, the WBMMBO takes care of the Feature Selection (FS) process. Subsequently, the same process is executed for the Query Images (QI) as well. Subsequently, the similarity is computed between the chosen features of the QI and that of the inputted image utilizing MSSIM for retrieving the image. Lastly, theThreshold-centered Checking (TC) is employed to check the retrieved image. The tentative outcomesdisclose that the proposed work can attainnoteworthy precision in addition to recall rates with superior computational efficiency.