基于可见光遥感的苎麻种质资源冠层性状研究

无人机近空遥感技术可快速实时掌握农田信息,在农作物田间监测中发挥日益重要的作用。本研究使用无人机可见光遥感平台,获取苎麻冠层航拍图像,通过图像处理获得苎麻种质资源冠层图像特征值,结合各苎麻种质资源生长性状,研究26份苎麻种质资源冠层图像性状差异。结果表明,使用HSV色彩空阈值分割可有效将苎麻与土壤杂草分割;26份苎麻资源6个表型性状变异系数分布在11.00%~40.00%之间,多样性指数分布在1.08~1.58之间;26份苎麻资源15个冠层颜色、纹理性状变异系数分布在0.28%~48.09%之间,多样性指数分布在1.25~1.54之间,表明试验苎麻种质资源具有丰富变异和广泛多样性。15个冠层颜色纹理性状主成分分析得到2个主成分,累计贡献率达到95.10%,可有效反映各性状的主要信息。

Studies on Canopy Characters of Ramie Germplasm Resources Based on Visible Light Remote Sensing

Unmanned aerial vehicle(UAV)near-air remote sensing technology provides the accessibility to monitor the farmland in a rapid and real-time manner. By taking use of the UAV visible light remote sensing platform,here the aerial images of canopy using 26 ramie germplasms were generated and analyzed for the characteristic values using the image processing pipeline. The results showed that HSV color image segmentation can effectively recognize ramie from soil weeds. The variation coefficient at 6 phenotypic traits of 26 ramie resources was 11.00%-40.00%,and the diversity index was1.08-1.58. The variation coefficients of 15 canopy color and texture traits of 26 ramie resources were distributed between 0.28% and 48.09%,and the diversity index was ranged from 1.25 to 1.54. That indicated a broad phenotypic variation in the tested ramie germplasm resources. Two principal components were identified by principal component analysis of 15 canopy color and texture traits,and the cumulative contribution rate reached 95.10%,which can effectively reflect the main information of each trait.