Abstract: Plant phenotypic traits are the commonly concerned data in genetic breeding and agricultural production. The traditional phenotype observation has become a major factor limiting the improvement of breeding efficiency and growth management. Plant phenotype detection based on visible image has advantages of low cost, heigh efficiency and easy popularization, and can be used not only to capture image of growing plants in suit, real-time and continuously, but also to extract various phenotype parameters, such as structure, shape, color and texture, etc.. This paper discussed the phenotype detection method and technology based on visible image for maize population, single maize plant and maize organs; introduced emphatically the progress in high-thoughput phenotype detection of maize root, leaf and ear. At the same time, the paper also analyzed the existing problems and development tendency, so as to provide references in formulating technical scheme and implementation procedures of crop phenotypic detection, and systematically carrying out studies on plant phenomics.

Key words: maize plant, phenotype detection, visible image, image processing

摘要： 植株表型性状是遗传育种和农业生产中普遍关心的数据，传统表型观测已成为限制育种效率和管理生长的主要因素。基于可见光图像的植株表型检测具有成本低、效率高、易推广等优点，不但可以原位、实时和连续获取植物生长图像，而且能够解析出植物结构、形态、颜色和纹理等多种表型参数。从玉米群体、单株和器官尺度上论述了基于可见光图像的玉米植物表型检测方法和技术，重点介绍了玉米根、叶、穗部等主要器官的高通量表型检测进展，并分析了基于可见光图像进行植物表型检测存在的问题和发展趋势，以期为制定作物表型检测技术方案和实施规程，及系统开展植物表型组学研究提供参考。

关键词: 玉米植株, 表型检测, 可见光图像, 图像处理