Scholarlink

<p>Accurate and geospatial estimation of fruit yield in densely structured orchards remains a major challenge for precision horticulture, primarily due to canopy heterogeneity, spectral saturation, and the limited physiological interpretability of conventional vegetation indices. This study proposes a novel, biophysically grounded framework for pixel-level yield estimation in kinnow mandarin (<i>Citrus reticulata</i>) orchards by integrating Sentinel-2–derived Red-Edge Position (S2REP) with key canopy traits, namely Leaf Area Index (LAI) and chlorophyll content. Field measurements were conducted across 550 individual tree canopies aggregated into 55 independent orchard pixels (20 × 20&#xa0;m), enabling direct linkage between satellite observations and in situ biophysical and yield data. Strong and statistically significant relationships were established between S2REP and both LAI (adjusted R² = 0.86, <i>p</i> &lt; 0.001) and chlorophyll content (adjusted R² = 0.80, <i>p</i> &lt; 0.001), confirming the sensitivity of red-edge signals to canopy structural and biochemical variations. These S2REP-derived traits were subsequently integrated into a multiple linear regression model for yield prediction, achieving high explanatory power (adjusted R² = 0.85) and robust generalization performance on independent validation data (validation R² = 0.82; RMSE = 46&#xa0;kg pixel⁻¹; MAE = 42&#xa0;kg pixel⁻¹). Predicted yields ranged from approximately 120 to 660&#xa0;kg per 20 × 20&#xa0;m pixel, with the majority of values predominantly concentrated between 300 and 500&#xa0;kg. Geospatial maps of LAI, chlorophyll content, and yield revealed pronounced intra-orchard variability, reflecting differences in canopy vigor, physiological status, and management conditions. By explicitly linking satellite-derived spectral features to physiologically meaningful canopy traits, the proposed framework enhances transparency, robustness, and scalability compared to purely empirical approaches. This methodology offers a cost-effective and operational solution for orchard-scale monitoring, early yield forecasting, and site-specific management, with direct implications for growers, policymakers, insurers, and agri-business stakeholders.</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

Fruit yield estimation of kinnow mandarin (Citrus reticulata) orchards – integrating canopy physiology with remote sensing

  • 期刊论文

  • 发表日期：

    2026年3月3日

查看全文
  • Ansar Ali,
  • Muhammad Imran,
  • Touqeer Ahmad,
  • Ghulam Abbas

摘要

Accurate and geospatial estimation of fruit yield in densely structured orchards remains a major challenge for precision horticulture, primarily due to canopy heterogeneity, spectral saturation, and the limited physiological interpretability of conventional vegetation indices. This study proposes a novel, biophysically grounded framework for pixel-level yield estimation in kinnow mandarin (Citrus reticulata) orchards by integrating Sentinel-2–derived Red-Edge Position (S2REP) with key canopy traits, namely Leaf Area Index (LAI) and chlorophyll content. Field measurements were conducted across 550 individual tree canopies aggregated into 55 independent orchard pixels (20 × 20 m), enabling direct linkage between satellite observations and in situ biophysical and yield data. Strong and statistically significant relationships were established between S2REP and both LAI (adjusted R² = 0.86, p < 0.001) and chlorophyll content (adjusted R² = 0.80, p < 0.001), confirming the sensitivity of red-edge signals to canopy structural and biochemical variations. These S2REP-derived traits were subsequently integrated into a multiple linear regression model for yield prediction, achieving high explanatory power (adjusted R² = 0.85) and robust generalization performance on independent validation data (validation R² = 0.82; RMSE = 46 kg pixel⁻¹; MAE = 42 kg pixel⁻¹). Predicted yields ranged from approximately 120 to 660 kg per 20 × 20 m pixel, with the majority of values predominantly concentrated between 300 and 500 kg. Geospatial maps of LAI, chlorophyll content, and yield revealed pronounced intra-orchard variability, reflecting differences in canopy vigor, physiological status, and management conditions. By explicitly linking satellite-derived spectral features to physiologically meaningful canopy traits, the proposed framework enhances transparency, robustness, and scalability compared to purely empirical approaches. This methodology offers a cost-effective and operational solution for orchard-scale monitoring, early yield forecasting, and site-specific management, with direct implications for growers, policymakers, insurers, and agri-business stakeholders.