Phenotyping earthworm by image analysis

Non-destructive phenotyping of earthworms by digital imaging and image analysis is the novel concept being proposed and explored in this paper. Earthworms are very important component of plant soil interaction via rhizosphere. Although a lot of research resources have been applied to phenotying roots by image analysis, there has been practically insignificant work on phenotying earthworms by image analysis. We put together some tailor made image analysis techniques (segmentation, medial axis thinning) along with a mathematical model for earthworms, to compute the volume, surface area and length of earthworms. We developed a novel radius versus length plot to identify the mouth-end, clitellum, and anus-end of earthworms by machine vision. We then compare the results of the phenotyping measurement obtained by our approach to those of the intercept principle. Intercept principle has been commonly used for phenotyping roots. Further more we propose a novel colour signature for blobs obtained by segmenting earthworms for colour analysis of the earthworms. It is expected that the colour information of earthworms can give clues on bioavailability of nutrients in soil or/and for earthworm species recognition. Both by qualitative and quantitative analysis we show that the segmentation and phenotype computation are better than the conventional approach of intercept principle.