Automated Spore Counting using Morphology and Shape

Abstract

Pyricularia Oryzae is a type of fungal spores which can lead to the most damaging rice blast disease. We have developed a quick and robust tool for counting the number of spores for measuring spore concentration using image processing techniques. The image is first thresholded using auto-Otsu's thresholding and adaptive Gaussian threshold. Morphological operations are employed to reduce some noise. With elongated shape of the spore, region properties are considered in the counting process. Our proposed technique is evaluated on 10x and 40x image sets using statistical measures; it outperforms the previous techniques and can be used for early disease diagnosis and further studying spore-related factors.