Plant Phenotyping : Picture this with machine vision

Plant phenotyping, which connects attributes of plant anatomy, physiology and performance back to their genetic origins and xenobiotic influences, is crucial for plant breeding. Traditional human observation is tarred with its subjective nature, drift over time, differences between observers and its often qualitative output. Machine vision offers solutions to these problems and in addition benefits from a fully quantitative output, abilities to look beyond human spectral perception, and measure parameters that are more challenging to the human observer. Image analysis coupled with high through-put automated systems is revolutionising plant phenotyping to remove a previous bottle neck to enable the rapid selection of favourable traits for plant breeders. This talk will focus on different plant structures, namely roots, aerial stems and leaves, fruit and seed, and how the software and hardware surrounding machine vision can be used to measure their respective phenotype parameters. A key to the success of image analysis for plant phenotyping within Syngenta has been to create a network of interdisciplinarv colleagues bound together by a common interest in imaging technologies and processing.

Download PDF