AUTHOR=Alshari Eman A. , Abdulkareem Mohammed B. , Gawali Bharti W. 

TITLE=Classification of land use/land cover using artificial intelligence (ANN-RF)

JOURNAL=Frontiers in Artificial Intelligence

VOLUME=Volume 5 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2022.964279

DOI=10.3389/frai.2022.964279

ISSN=2624-8212

ABSTRACT=Because deep learning has various downsides, such as complexity, expense, and the need to wait longer for results, this produces a significant incentive and impetus to invent and adopt this notion to develop machine learning because it is simple. This study intended to increase the accuracy of machine learning approaches for land use/land cover classification using Sentinel-2A -2A and landsat8 satellites. This research aims to implement a proposed method, neural-based with objected-based, to produce a model addressed by Artificial Neural Networks (limited parameters) with Random Forest (Hyper-parameter) called (ANN_RF). This study used multispectral satellite images (Sentinel-2A & landsat8) and a normalized digital elevation model as input datasets for the Sana'a city map of 2016. The results showed that the accuracy for the proposed model (ANN_RF) is better than the ANN classifier with the Sentinel-2A and landsat8 Satellites individually, which may contribute to developing machine learning through newer researchers and specialists. The proposed work develops traditional artificial neural networks with seven to ten layers in a conventional way, but now with access to thousands and millions of simulated neurons without deep learning techniques by this proposed approach (ANN_RF). In this study, the neural networks typically consist of five input and three hidden layers. The output is for one class, there are six basic parameters, and each parameter has multi courses and dozens of neurons.