Evaluation of symptoms on a plant is a simple mean of determining the progress of disease. For its speed and unnecessity of special equipment is the visual evaluation still a widely used method. It assesses disease symptoms by various scales and guides. But due to individual subjective evaluation this method does not fulfil requirements on accuracy and reproducibility. These deficits can be partially or fully eliminated by using methods of image processing.
Image processing has a wide application in examination of plant-pathogen interactions, mainly in combination with various laboratory methods like specific colouring of the pathogen body in leaves or colouring of plant’s defensive structures and substances and their co-localization with pathogen.
A specialized image-processing tool DPlan was created to enable semi-automatic processing of large series of images. The tools is algorithmically based on library ITK and is developed using other open-source and multi-platform tools Qt, CMake and NSIS. DPlan is distributed under GPL license (see the web pagehttp://4lab.smyslzivota.cz/DPlan/).
Several experiments were presented, where pathogen body or plant’s defensive structures were coloured, the leaves were scanned, and resultant digital image was analyzed by DPlan.
A. Leaf of rape with lesion. B. The same leaf coloured by trypan blue. C. The result of image segmentation: pink colour highlights the background and green colour marks the detected lesion.
|Contact person:||Jakub Bican|
|Involved extern:||Vladimír Šašek, Lenka Burketová (UEB AV ČR)|