Remote Sensing-Based Earth Climate Detection in Geoscience Model with Artificial Intelligence Application

• Published in: Remote sensing in earth systems sciences (Online) 2024-12, Vol.7 (4), p.569-581
• Main Authors: Agarkar, Aarti Amod, Karyakarte, Mandar S., Chavhan, Gajanan, Ariffin, I. A., Patil, Milind, Ushasree, Linginedi, Priya, D. Divya
• Format: Article
• Language: English
• Subjects: Earth and Environmental Science; Earth System Sciences; Geography; Monitoring/Environmental Analysis; Remote Sensing/Photogrammetry
• ISSN: 2520-8195; 2520-8209
• DOI: 10.1007/s41976-024-00146-8

Remote sensing technology can offer exact information in an image format, it has permeated every sector related to natural resources. Right now, the source of geographic area data that is expanding the fastest is satellites used for remote sensing. Change detection is a highly valuable tool for monitoring environmental and human demands, given the constant changes in the earth’s surface and the extensive use of remote sensing. A significant capacity for estimating climate change has been made possible by combining machine learning (ML) models with remote sensing imagery. This methodology takes less data than traditional methods and relaxes a number of assumptions. Influential variables for estimating climate change using machine learning models are provided by satellite imagery. In this research, climate change detection has been analysed for geoscience modelling based on remote sensing data and ML models. Remote-sensed region images have been collected in which the pre-processing is carried out for noise removal, normalisation, and smoothening. This processed image’s features have been selected using a linear principal recurrent component neural network and classified using a random graph SqueezeNet perceptron neural network. The simulation results have been analysed for various remote-sensed region image data in terms of random accuracy, average precision, RMSE, recall, and F -measure. The proposed technique has a random accuracy of 96%, average precision of 97%, F -measure of 87%, RECALL of 85%, and RMSE of 61%.