Scholarlink

<p>In India, mustard (Brassica juncea) is a significant cash and oilseed crop. This work aims to evaluate the different mustard crop classification algorithms for mapping mustard crop in Bharatpur District, Rajasthan, India, using Google’s Earth Engine (GEE) cloud platform and integrates socio-economic, economic, and terrain-topography parameters into the crop suitability analysis. Several methods were used, including unsupervised and supervised classification, HDRBC, RF, and machine learning-based classification, over multi-temporal satellite data captured 15 days apart from October 2022 until March 2023-Rabi, and geospatial indicators generated include FCC, NDVI, land use/land cover crop masks, and time series vegetation dynamics through custom GEE codes. According to the government statistics report, the area under mustard cultivation for the 2022–2023 season was around 274,500&#xa0;ha. According to the model predictions, the areas under mustard cultivation are unsupervised, 296,763&#xa0;ha; supervised, 290,954&#xa0;ha; HDRBC, 280,534&#xa0;ha; and RF, 278,541&#xa0;ha. The model with the best accuracy (93.67%) was the RF, followed by HDRBC (87.20%), supervised (79.24%), and unsupervised (67.65%). To further ensure sustainable expansion in production, a crop suitability analysis of 20 parameters, 10 of which were socio-economic and 10 economic, was integrated with terrain and topographic parameters. The suitability map generated classified the study area into very low (9.27%), low (18.74%), moderate (27.27%), high (19.60%), and very high (25.11%) suitability areas. The results revealed a region such as Bharatpur, Sikri Ratti, Nagar, Semli, Pahari, Gopalgarh, Rarah, Chak Sana, and Jagiria, where mustard cultivation is currently supported under less favorable conditions. Pathena, Halena, Bhasawar, Wer, Karsado, Nithor, Luharu, Uchen, Khanua, and Kurka are areas exhibiting high to very high suitability. This study establishes that integrating GEE-based ML crop mapping with socio-economic and terrain-based suitability analysis effectively classifies optimal zones for mustard cultivation, supporting sustainable strategies to improve productivity.</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

A google earth engine–driven machine learning framework for mustard crop mapping and suitability index assessment

  • 期刊论文

  • 发表日期：

    2026年6月18日

查看全文
  • Ahamed Ibrahim Abdul Rahim,
  • Prabhakaran Moorthy,
  • Bharathi Balu,
  • Muthusankar Gowrappan,
  • Lakshumanan Chokkalingam

摘要

In India, mustard (Brassica juncea) is a significant cash and oilseed crop. This work aims to evaluate the different mustard crop classification algorithms for mapping mustard crop in Bharatpur District, Rajasthan, India, using Google’s Earth Engine (GEE) cloud platform and integrates socio-economic, economic, and terrain-topography parameters into the crop suitability analysis. Several methods were used, including unsupervised and supervised classification, HDRBC, RF, and machine learning-based classification, over multi-temporal satellite data captured 15 days apart from October 2022 until March 2023-Rabi, and geospatial indicators generated include FCC, NDVI, land use/land cover crop masks, and time series vegetation dynamics through custom GEE codes. According to the government statistics report, the area under mustard cultivation for the 2022–2023 season was around 274,500 ha. According to the model predictions, the areas under mustard cultivation are unsupervised, 296,763 ha; supervised, 290,954 ha; HDRBC, 280,534 ha; and RF, 278,541 ha. The model with the best accuracy (93.67%) was the RF, followed by HDRBC (87.20%), supervised (79.24%), and unsupervised (67.65%). To further ensure sustainable expansion in production, a crop suitability analysis of 20 parameters, 10 of which were socio-economic and 10 economic, was integrated with terrain and topographic parameters. The suitability map generated classified the study area into very low (9.27%), low (18.74%), moderate (27.27%), high (19.60%), and very high (25.11%) suitability areas. The results revealed a region such as Bharatpur, Sikri Ratti, Nagar, Semli, Pahari, Gopalgarh, Rarah, Chak Sana, and Jagiria, where mustard cultivation is currently supported under less favorable conditions. Pathena, Halena, Bhasawar, Wer, Karsado, Nithor, Luharu, Uchen, Khanua, and Kurka are areas exhibiting high to very high suitability. This study establishes that integrating GEE-based ML crop mapping with socio-economic and terrain-based suitability analysis effectively classifies optimal zones for mustard cultivation, supporting sustainable strategies to improve productivity.