Scholarlink

This study introduces a novel framework for forest fire detection and carbon emission estimation, employing a Gram-Schmidt Gauss-MishGoogleNet (GSGMGN). Utilizing remote sensing data from the Himalayan region of Sikkim, India, the framework enhances image contrast through Double Plateau Histogram Equalization (DPHE) and employs Haversine and Silhouette Density-Based Spatial Clustering of Applications with Noise (HSDBSCAN) to accurately delineate and cluster forest regions. The approach subsequently identifies densely forested areas and implements an Adaptive Gaussian Z-scored Neuro-Fuzzy Inference System (AGZNFIS) to perform comprehensive burn area assessments. Key vegetation features are extracted, with significant indices selected through a Guided Gannet Initialized Optimization Algorithm (G2IOA). These features feed into the GSGMGN model, which estimates carbon emissions based on burn areas, vegetation indices, biomass, and weather data. Comparative analysis with existing methods reveals that the GSGMGN framework achieves superior accuracy at 99.1%, with carbon emissions estimated at 1678 tons. This framework not only provides precise detection but also advances carbon emission estimation in forest fire-affected regions.

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

A Carbon Severity Level Assessment Utilizing GSGMGN and AGZNFIS Approaches in Sikkim Himalayas

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Kapila Sharma,
  • Gopal Thapa

摘要

This study introduces a novel framework for forest fire detection and carbon emission estimation, employing a Gram-Schmidt Gauss-MishGoogleNet (GSGMGN). Utilizing remote sensing data from the Himalayan region of Sikkim, India, the framework enhances image contrast through Double Plateau Histogram Equalization (DPHE) and employs Haversine and Silhouette Density-Based Spatial Clustering of Applications with Noise (HSDBSCAN) to accurately delineate and cluster forest regions. The approach subsequently identifies densely forested areas and implements an Adaptive Gaussian Z-scored Neuro-Fuzzy Inference System (AGZNFIS) to perform comprehensive burn area assessments. Key vegetation features are extracted, with significant indices selected through a Guided Gannet Initialized Optimization Algorithm (G2IOA). These features feed into the GSGMGN model, which estimates carbon emissions based on burn areas, vegetation indices, biomass, and weather data. Comparative analysis with existing methods reveals that the GSGMGN framework achieves superior accuracy at 99.1%, with carbon emissions estimated at 1678 tons. This framework not only provides precise detection but also advances carbon emission estimation in forest fire-affected regions.