Image-based field plant phenotyping approaches for modern agriculture

Abstract

Plant phenotyping becoming increasingly important in modern agriculture, it investigates how a plant's genome, interacting with the environment, affects the observable traits of a plant. To solve the destructive and labor-intensive limitations of phenotyping, new technique known as “image-based Phenotyping” is being conducted successfully under controlled environment such as breeding industry. However, in real agricultural field the plant phenotype is formed under the dynamic interaction between genotype and environment, phenotypes generated from controlled environments experiments do not always correlate well with typical field behavior of plants. Moreover, different from the situation of plants grown individually in pots under controlled environment, plants in the field do not grow isolated but instead are free to interact with neighboring plants, for example via their root systems. Therefore, phenotypic characteristics such as canopy configuration measured in plants communities grown under uncontrolled outdoor environments are different those of individual plants. Thus, there is a strong need to establish high-throughput phenotyping methods that can be used to screen crop populations under natural environmental conditions in the field. In this paper, we introduce several approaches that aimed to contribute to field phenotyping system that will be usable in natural environments, particularly based on RGB images collected by low cost field monitoring systems.