基于语义分割的大豆主茎精准表型自动获取及其在导入系材料选择中的应用

大豆导入系群体的构建和材料的选择是大豆育种的重要手段之一,而针对某一具体表型的材料选择不仅能够提高育种效率,而且还能够加快定位目标性状基因的进程。本研究针对由绥农14(轮回亲本)和野生野生大豆ZYD00006(供体亲本)为双亲构建的导入系(190个个体),采用机器视觉中的语义分割技术,对主茎相关表型进行了自动提取,实验结果表明:此方法完全可以取代人工测量且具有良好的可移植性,同样可用于其他作物的表型获取。在此基础上,本研究针对大豆主茎相关表型,对参试材料进行系统聚类,结果明确给出各参试材料(近等材料)基于主茎表型的聚类情况。这一结果将成为针对主茎表型进行材料选择的重要依据,同时依据这一结果可以大大加快主茎相关表型基因的定位。

Automatic acquisition of accurate phenotype from stems of soybean based on semantic segmentation and its application in selection of imported line materials

The construction of soybean imported line population and selection of materials are important in soy bean breeding.The selection of imported line materials for a specific phenotype can not only improve the efficiency of breeding,but also speed up the process of locating target trait genes.In this study,the semantic segmentation technique in machine detection was used to automatically extract the main stem-related phenotypes from the import ed lines(190 individuals)constructed for both parents,Suinong 14(recurrent parent)and wild soybean ZYD00006(donor parent).On this basis,this study systematically clustered the tested materials for the main stem-related phe notypes,and the results clearly showed the aggregation of the tested materials(near-equal materials)based on the main stem-phenotypes.It became an important basis for the selection of materials for the main stem-phenotypes.At the same time,the results could greatly accelerate the locating procedure of the main stem related phenotype genes.