| 摘要: |
| 倒伏作物产量评估是作物倒伏灾害综合评估中的一项重要指标。针对传统倒伏作物产量评估方法 的不足，该文提出了一种基于图像处理技术和光谱分析技术的倒伏冬小麦估产方法。研究首先通过冬小麦 倒伏模拟试验，人工造成冬小麦倒伏，采集不同倒伏等级冬小麦RGB图像数据、冠层光谱反射率数据，并 测量倒伏冬小麦产量。然后，通过RGB图像获得颜色、纹理特征，通过光谱反射率获得光谱吸收、红边和 植被指数特征，取与倒伏冬小麦产量最为相关的23个特征量作为产量预测模型的输入变量。最后，使用粒 子群算法优化径向基权值并建立估产模型，通过三个倒伏时期的光谱和图像指标估算倒伏冬小麦产量。结 果表明，模型估算精度为98.2%，该模型能满足非接触定量化的倒伏冬小麦产量评估需求。 |
| 关键词: 冬小麦 倒伏 图像 高光谱 估产 |
| DOI：10.7621/cjarrp.1005-9121.20130108 |
| 分类号: |
| 基金项目:该研究得到国家科技支撑计划（2012BAH29B00）资助 |
|
| STUDY ON YIELD ASSESSMENT OF LODGING WINTER WHEAT BASED ON IMAGE MANIPULATION AND HYPERSPECTRAL ANALYSIS |
|
Wu Shangrong1, Liu Jia2, Wang Limin2, Li Zongnan2
|
|
1.National Key Laboratory for Electronic Measurement Technology， North University of China， Taiyuan 030051， China;2.Institute of Agricultural Resources and Regional Planning， Chinese Academy of Agricultural Sciences， Beijing 100081， China/Key Laboratory of Agri-informatics， Ministry of Agriculture， Beijing 100081， China
|
| Abstract: |
| Lodging crop yield assessment is an important indicator for the comprehensive evaluation of crop lodging disasters. Aiming at the shortage of the traditional methods of lodging crop yield assessment， the paper proposed a method of estimating lodging winter wheat yield based on image manipulation and hyperspectral analysis. Firstly， the study designed a winter wheat lodging experiment and captured different grades of lodging winter wheat RGB im- age data， canopy spectral reflectance data， and the yield of winter wheat. Secondly， by analyzing the RGB image color and texture characteristics， spectral absorption features red-edge characteristics and vegetation index， it se- lected 23 features mostly related to winter wheat yield as the input variables of yield assessment model. Finally， it optimized radial basis function（RBF） weights and established yield estimation model using particle swarm optimi- zation（PSO）， the results indicated that the prediction accuracy of PSO-RBF model was 98.2%， the method based on hyper-spectral analysis and image manipulation to estimate the lodging winter wheat yield can meet the needs of non-contact quantitative assessment. |
| Key words: winter wheat lodging image hyper spectral yield estimation |
