Remotely Sensed Image Classification Using Deep Learning Algorithm with Bio Inspired Optimizers

Abstract

Remote sensing has become pivotal in various fields for monitoring and analyzing Earth's surface. Deep learning algorithms, particularly AlexNet, have shown promise in image classification tasks. However, effective feature extraction remains a challenge. Bio-inspired optimizers like Bee Swarm Optimization (BSO) offer a promising solution for enhancing feature extraction. This study proposes a novel approach combining AlexNet deep learning for image classification with BSO for feature extraction from remote sensed image data. The AlexNet architecture is utilized for its robustness in handling complex image data, while BSO optimizes feature extraction to enhance classification accuracy. The AlexNet with BSO presents a unique approach to improving remote sensed image classification accuracy. Experimental results demonstrate that the combined approach outperforms traditional methods in remote sensed image classification tasks. The utilization of BSO for feature extraction enhances the discriminative power of the model, leading to improved accuracy and robustness.