Hybrid deep learning with pelican optimization algorithm for M2M communication on UAV image classification

Abstract

Machine-to-machine (M2M) communication for unmanned aerial vehicle (UAVs) and image classification is essential to current remote sensing and data processing. UAVs and ground stations or other linked devices exchange information seamlessly using M2M communication. M2M connectivity helps UAVs with cameras and sensors communicate aerial pictures in real time or post-mission for image categorization and analysis. During flight, UAVs acquire massive volumes of picture data. Image classification, commonly using deep learning (DL) methods like convolutional neural network (CNN), automatically categorizes and annotates photos based on predetermined classes or attributes. This work uses UAV photos to produce hybrid deep learning with pelican optimization algorithm for M2M communication (HDLPOA-M2MC). HDLPOA-M2MC automates UAV picture class identification. GhostNet model is used to derive features in HDLPOA-M2MC. The HDLPOA-M2MC approach leverages pelican optimization algorithm (POA) for hyperparameter adjustment in this investigation. Finally, autoencoder-deep belief network (AE-DBN) model can classify. The HDLPOA-M2MC method’s enhanced outcomes were shown by several studies. The complete results showed that HDLPOA M2MC performed better across measures.