Title: Application of U-Net remote sensing data in ecological landscape restoration planning and pollution prevention

Authors: Jiaxin Li; Bonhak Koo

Addresses: Department of Green Smart, Graduate School, Sangmyung University, Cheonan, 31066, South Korea ' Department of Green Smart City (GSC), Sangmyung University, Cheonan, 31066, South Korea

Abstract: This paper uses the U-Net model to accurately analyse remote sensing data and applies it to ecological landscape restoration planning and pollution prevention to improve environmental monitoring efficiency, optimise restoration strategies, and help achieve sustainable management of ecosystems. First, a large amount of remote sensing image data is obtained through Landsat satellite images and preprocessed. Then, the U-Net model is used to analyse the remote sensing data, identify different types of land objects, and monitor pollution sources in the environment. Autoencoder is used to detect abnormal areas in the land object classification results, and finally a support vector machine is used to classify the pollution sources. The results show that the average accuracy of the U-Net model in classifying landscape features reaches 95.4%, and the use of Autoencoder can accurately detect abnormal areas; the combination of U-Net and remote sensing data can achieve accurate classification of features.

Keywords: ecological landscape; restoration planning; pollution prevention; remote sensing data; U-Net model; autoencoder model.

DOI: 10.1504/IJEP.2024.143454

International Journal of Environment and Pollution, 2024 Vol.75 No.1, pp.1 - 20

Received: 29 Aug 2024
Accepted: 03 Oct 2024
Published online: 20 Dec 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article