Enhanced Satellite Image Classification Using Deep Convolutional Neural Network
Bihari Nandan Pandey, Mahima Shanker Pandey · 2024
Satellite image classification techniques involve numerous approaches from segmentation to classification using nature inspired algorithms, swarm intelligence approaches and now using different supervised and unsupervised learnings algorithms. This classification involves complicated task as images contains different types of imagery, and different boundaries are overlapped to each other which must be separated so that the useful information must be extracted from the satellite images. In this paper we propose an innovative approach for satellite image classification by integrating Deep Convolutional Neural Network (CNN). The CNN architecture offers a solution by combining the strengths of Convolutional Neural Networks (CNN) in feature extraction with Deep Neural Networks in capturing temporal dependencies. Through this fusion, the model gains enhanced capability in discerning complex patterns within satellite images. The proposed framework undergoes rigorous evaluation on the image set, achieving an remarkable accuracy of 98% in classifying satellite images, surpassing benchmarks set by conventional methods. Such high accuracy makes it a promising tool for various applications, including land use monitoring, disaster management, and urban planning. The findings underscore the potential of CNN-based approaches in advancing satellite image analysis, paving the way for more accurate and efficient remote sensing technologies.