Abstract
A method to obtain clear binary images of bread crumb grain has been developed from the viewpoint of image synthesis, and named ‘Multiple Binarization Imaging Method (MBI).’ Nine thresholds were formally selected from a brightness histogram with 256 gray-levels. The object image was synthesized (logical add) from nine intermediate binary images with each threshold after eliminating unnecessary image element. The aim of this study is to examine the MBI method using bread-crumb simulated boards which had holes with constant diameter (3.0mm) and various depths (0.5-5.0mm), and those with constant depth (3.0mm) and various diameters (0.5-5.0mm). Brightness value of a hole is considered to be correlated to the depth of the hole. However, the hole diameter could not be identified from brightness value histogram with 256 gray-levels. Therefore, it was suggested that the number of thresholds which should be set did not depend on the hole diameter and the number of patterns of hole depths, but that it was derived from the fact that the gray image of air hole images had 10 classes. The MBI method yielded good and consistent binary images of crumb air bubbles of commercial white bread. Open-top type bread had coarser crumb structure with vertically orientated at the central area of the crumb. Pullman-type bread had finer porous structure without orientation at the central area of the crumb.
