Classification of Satellite Images Using a Deep Learning-Inspired Hybrid Novel Approach

Satellite imagery is crucial for disaster assistance, law enforcement, and environmental monitoring.Some users need to identify facilities and items in photographs manually.Automation becomes essential when there are large areas to search and few available analysts.However, the problem can only be fixed by increasing the precision of existing object identification and categorization methods.The "deep learning" subfield of machine learning has demonstrated promising results for automating specific tasks.Using convolutional neural networks, it was able to understand images successfully.In this work, we use high-resolution, multi-spectral satellite photos to solve the problem of identifying objects and infrastructure.In this paper, we describe a deep-learning system for labeling objects.In this research, we make use of the Satellite Image Classification Dataset-RSI-CB256.This dataset uses Google Maps images and sensors to create four distinct categories.In this study, a hybrid model is proposed, which achieves an accuracy of 98.96%.