Scholarlink

<p>This study presents a comprehensive approach for mapping harvested groundnut fields using advanced remote sensing techniques and a&#xa0;fuzzy classification model. Using temporal data from two sensors, including PlanetScope and Sentinel-1 Synthetic Aperture Radar, a dual sensor approach was employed to accurately identify groundnut harvesting dates and map crop fields with precision. The Adaptive Modified Possibilistic Local Information <i>c</i>-Means (ADMPLICM) algorithm for incorporating contextual information along with the training approach as ‘Individual sample as mean’ (ISM) emerged as a vital tool in efficiently handling the heterogeneity and noisy pixels of groundnut fields. Its adaptive nature, incorporating spatial attraction between pixels and local similarity measures, facilitated robust classification, even in the presence of varying field characteristics and environmental conditions. The study estimated the total harvested groundnut area to be 23.4 square kilometers, providing valuable insights for agricultural planning and resource allocation. Through temporal trend analysis, distinct patterns in groundnut harvesting activity were revealed, highlighting the importance of seasonality in crop management. The average Mean Membership Difference was calculated to be 0.00379, with a Variance of 0.00003 within the target crop, demonstrating the accuracy achieved using the ADMPLICM ISM approach. This interdisciplinary approach bridges remote sensing, fuzzy classification, and agricultural science, offering a reliable and efficient method for mapping harvested groundnut fields. The findings contribute to informed decision-making and sustainable agricultural management practices, with implications for precision agriculture and remote sensing applications across various crop types and geographical regions.</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

Temporal Harvest Mapping of Groundnut Crop: A Contextual Fuzzy Machine Learning Approach

  • 期刊论文

  • 发表日期：

    2025年6月21日

查看全文
  • Suhas Muralidhar,
  • Anil Kumar

摘要

This study presents a comprehensive approach for mapping harvested groundnut fields using advanced remote sensing techniques and a fuzzy classification model. Using temporal data from two sensors, including PlanetScope and Sentinel-1 Synthetic Aperture Radar, a dual sensor approach was employed to accurately identify groundnut harvesting dates and map crop fields with precision. The Adaptive Modified Possibilistic Local Information c-Means (ADMPLICM) algorithm for incorporating contextual information along with the training approach as ‘Individual sample as mean’ (ISM) emerged as a vital tool in efficiently handling the heterogeneity and noisy pixels of groundnut fields. Its adaptive nature, incorporating spatial attraction between pixels and local similarity measures, facilitated robust classification, even in the presence of varying field characteristics and environmental conditions. The study estimated the total harvested groundnut area to be 23.4 square kilometers, providing valuable insights for agricultural planning and resource allocation. Through temporal trend analysis, distinct patterns in groundnut harvesting activity were revealed, highlighting the importance of seasonality in crop management. The average Mean Membership Difference was calculated to be 0.00379, with a Variance of 0.00003 within the target crop, demonstrating the accuracy achieved using the ADMPLICM ISM approach. This interdisciplinary approach bridges remote sensing, fuzzy classification, and agricultural science, offering a reliable and efficient method for mapping harvested groundnut fields. The findings contribute to informed decision-making and sustainable agricultural management practices, with implications for precision agriculture and remote sensing applications across various crop types and geographical regions.