Plant Phenotyping with Low Cost Digital Cameras and Image Analytics

摘要

In this paper we discuss a prototype, easy-to-deploy, and low cost (~ $250) phenotype collection system for growth chambers. Off the shelf digital cameras, wireless transmitters, and PCs are used to store and process the images. A Matlab pipeline is used to fuse multiple images, identify the location of each Arabidopsis plant, segment its leaves, and measure leaf topology and area. Our early findings (unpublished) using this system for correlating genotype to phenotype are very promising. We hope that with future improvements and broad adoption, it will have the same disruptive effects as the first "build your own" microarrayers, which allowed for the explosion of genotyping information. Low cost genotyping and phenotyping will hopefully address some of the environmental, agricultural, and industrial sustainability challenges we are facing today.