MAPPING AGRICULTURAL WASTE FOR BIOGAS PRODUCTION USING A FULLY CONVOLUTIONAL NEURAL NETWORK AND REMOTE SENSING IMAGERY

Abstract

The increasing generation of waste and depletion of natural resources has led to a growing need for innovative approaches for utilizing different types of waste as potential energy and material resources. Agricultural activities produce large amounts of Agricultural Waste (AW), which, if not adequately managed, can lead to environmental degradation. One potential solution for the effective utilization of AW is converting it into biogas. However, commercializing this process requires a comprehensive understanding of the types and quantities of AW generated. In this paper, the use of a Fully Convolutional Neural Network (FCN), which has rapidly advanced with the progress of Artificial Intelligence and become essential for tasks such as Semantic Segmentation, Object Detection, and Image Classification, is proposed to improve the prediction of AW for biogas production. Furthermore, this paper presents a Deep Learning-based image segmentation method to recognize vineyard fields, which are a significant source of AW, using remote satellite images. The proposed approach can significantly improve the identification of AW sources, and thus contribute to the efficient and sustainable utilization of AW for biogas production.