Scholarlink

<p>Forests contribute significantly to climate regulation, biodiversity conservation, and the livelihoods of billions of people. The advancement of digital technologies has driven the growing use of Earth observation techniques, involving satellites, drones, and sensors to gather large volumes of heterogeneous environmental data. However, integrating and interpreting these diverse data sources, such as vegetation indices, meteorological records, geospatial information, and sensor measurements, remains a significant challenge. This paper presents a framework that constructs an integrated knowledge base built around a custom ontology, <i>SORSOntology</i>, designed to provide a consistent and accessible representation of environmental information. By combining AI-driven image analysis with semantic technologies, the proposed system enables automated deforestation monitoring and contextual reasoning driven by both OWL class restrictions and SWRL rules. These inference mechanisms support the automatic classification of observations into high-level environmental categories, and are transparently executed in the backend upon image selection within the Web-based user interface. Experimental results on Sentinel-2 imagery of the Amazon region show that the framework achieves accurate segmentation performance and allows enriched interpretation of forest conditions through ontology-based queries.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Constructing a knowledge base from remote sensing indicators for deforestation assessment

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月10日

查看全文
  • Giacomo Albamonte,
  • Giorgio Falcone,
  • Manilo Monaco,
  • Sabrina Senatore

摘要

Forests contribute significantly to climate regulation, biodiversity conservation, and the livelihoods of billions of people. The advancement of digital technologies has driven the growing use of Earth observation techniques, involving satellites, drones, and sensors to gather large volumes of heterogeneous environmental data. However, integrating and interpreting these diverse data sources, such as vegetation indices, meteorological records, geospatial information, and sensor measurements, remains a significant challenge. This paper presents a framework that constructs an integrated knowledge base built around a custom ontology, SORSOntology, designed to provide a consistent and accessible representation of environmental information. By combining AI-driven image analysis with semantic technologies, the proposed system enables automated deforestation monitoring and contextual reasoning driven by both OWL class restrictions and SWRL rules. These inference mechanisms support the automatic classification of observations into high-level environmental categories, and are transparently executed in the backend upon image selection within the Web-based user interface. Experimental results on Sentinel-2 imagery of the Amazon region show that the framework achieves accurate segmentation performance and allows enriched interpretation of forest conditions through ontology-based queries.