Scholarlink

<p>Rural livelihood strategies significantly influence socio-ecological systems and sustainability outcomes. However, spatially explicit data on livelihood compositions remain scarce at national scales. We developed a Deep Rural Livelihood Model (DRLM) combining satellite imagery (Landsat, VIIRS) with household survey data to map four livelihood strategy probabilities across rural China in 2020: farming-only (F), farming-dominated mixed (F_NF), non-farming-dominated mixed (NF_F), and non-farming-only (NF). Using quantile-regression XGBoost, we expanded survey-derived labels from 283 rural observation sites to 38,306 rural settlements and trained a dual-branch ResNet architecture with Dirichlet regression. Agreement with the expanded settlement-level labels was high, whereas independent validation against actual survey data showed moderate performance, with stronger support for dominant livelihood categories than for mixed categories. This dataset provides settlement-scale probabilistic information to support analyses of rural transformation, sustainability-related indicators, and socio-ecological change.</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

Mapping rural livelihood strategies in China using deep learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月21日

查看全文
  • Zhaxi Dawa,
  • Wenjuan Yu,
  • Weiqi Zhou

摘要

Rural livelihood strategies significantly influence socio-ecological systems and sustainability outcomes. However, spatially explicit data on livelihood compositions remain scarce at national scales. We developed a Deep Rural Livelihood Model (DRLM) combining satellite imagery (Landsat, VIIRS) with household survey data to map four livelihood strategy probabilities across rural China in 2020: farming-only (F), farming-dominated mixed (F_NF), non-farming-dominated mixed (NF_F), and non-farming-only (NF). Using quantile-regression XGBoost, we expanded survey-derived labels from 283 rural observation sites to 38,306 rural settlements and trained a dual-branch ResNet architecture with Dirichlet regression. Agreement with the expanded settlement-level labels was high, whereas independent validation against actual survey data showed moderate performance, with stronger support for dominant livelihood categories than for mixed categories. This dataset provides settlement-scale probabilistic information to support analyses of rural transformation, sustainability-related indicators, and socio-ecological change.