Journal of Integrative Agriculture ›› 2017, Vol. 16 ›› Issue (07): 1474-1485.DOI: 10.1016/S2095-3119(16)61542-3

• 论文 • 上一篇    下一篇

  

  • 收稿日期:2016-07-28 出版日期:2017-07-20 发布日期:2017-07-06

Estimating light interception using the color attributes of digital images of cotton canopies

XUE Hui-yun1, 2, HAN Ying-chun1, LI Ya-bing1, WANG Guo-ping1, FENG Lu1, FAN Zheng-yi1, DU Wen-li1, YANG Bei-fang1, MAO Shu-chun1    

  1. 1 Institute of Cotton Research, Chinese Academy of Agricultural Sciences/State Key Laboratory of Cotton Biology, Anyang 455004, P.R.China
    2 Henan Collaborative Innovation Center of Modern Biological Breeding, Henan Institute of Science and Technology, Xinxiang 453003, P.R.China
  • Received:2016-07-28 Online:2017-07-20 Published:2017-07-06
  • Contact: Correspondence LI Ya-bing, Tel: +86-372-2562293, E-mail: criliyabing1@163.com; MAO Shu-chun, Tel: +86-372-2562216, E-mail: maosc@cricaas.com.cn
  • About author:XUE Hui-yun, E-mail: xuehy8310@163.com;
  • Supported by:

    This study was supported by the National Natural Science Foundation (31371561). We gratefully acknowledge the help of technicians from the experimental station of the Institute of Cotton Research of the Chinese Academy of Agricultural Sciences.

Abstract:     Crop growth and yield depend on canopy light interception (LI). To identify a low-cost and relatively efficient index for measuring LI, several color attributes of red-green-blue (RGB), hue-saturation-intensity (HSI), hue-saturation-value (HSV) color models and the component values of color attributes in the RGB color model were investigated using digital images at six cotton plant population densities in 2012–2014. The results showed that the LI values followed downward quadratic curves after planting. The red (R), green (G) and blue (B) values varied greatly over the years, in accordance with Cai’s research demonstrating that the RGB model is affected by outside light. Quadratic curves were fit to these color attributes at six plant population densities. Additionally, linear regressions of LI on every color attribute revealed that the hue (H) values in HSI and HSV were significantly linearly correlated with LI with a determination coefficient (R2)≥0.89 and a root mean square error (RMSE)=0.05. Thus, the H values in the HSI and HSV models could be used to measure LI, and this hypothesis was validated. The H values are new indexes for quantitatively estimating the LI of heterogeneous crop canopies, which will provide a theoretical basis for optimizing the crop canopy structure. However, further research should be conducted in other crops and under other growing and environmental conditions to verify this finding.

Key words: canopy light interception (LI) ,  digital image ,  color attributes ,  hue