Abstract
A computer-based image processing technique was adapted for rapid screening of Bacillus thuringiensis. Sporulated colonies of the B. thuringiensis/B. cereus group were observed under a phase-contrast microscope and the images were recorded by a CCD camera. The data were converted into binary images after reducing the background noises and thickening the edge lines of spores and inclusions. Parasporal inclusions were then distinguished from spores by a difference in curvatures and sizes. The system automatically differentiated in five seconds between B. thuringiensis and B. cereus by the ratio of spores and parasporal inclusions.
