TY - JOUR
TI - Semantic segmentation for mapping agricultural waste sources: A vineyard case study for energy valorization via biogas production
AU - Petrović Emina P
AU - Momčilović Ana J
AU - Dimitrijević-Jovanović Dragana
AU - Stefanović Gordana M
AU - Simonović Miloš B
AU - Milošević Maša
AU - Nikolić Vlastimir D
JN - Thermal Science
PY - 2025
VL - 29
IS - 5
SP - 3319
EP - 3329
PT - Article
AB - Given the growing trend of increasing waste and diminishing resources, considerable efforts are being directed toward developing innovative methods for utilizing various types of waste as potential energy and material resources. Agriculture generates large quantities of waste, and inadequate management of this waste can cause severe environmental challenges. Transforming agricultural waste into biogas presents an excellent opportunity for its effective use, however, commercializing this process requires a comprehensive understanding of potential agricultural waste sources, primarily the types and quantities of waste generated. Consequently, this paper proposes a deep learning-based image segmentation approach for identifying potential agricultural waste sources using remote sensing images. The research examines the effectiveness of the DeepLabV3+ with various backbone networks for semantic segmentation with an emphasis on detecting vineyards as potential contributors to agricultural waste for biogas generation.