Measurement of bread cell structure by image analysis.

ABSTRACT A method of measuring the cell structure of bread by digital image analysis is described. This provides an objective alternative to visual assessments as a tool for quality control and for critical evaluation of process and ingredient variations. The method is also applicable to other baked products with a cellular structure. Loaves were sliced in a consistent manner and placed in a specially constructed, light-proof cabinet incorporating illumination and a CCD camera. Oblique illumination was used to accentuate the cell structure at the slice surface. The brightness scale was calibrated automatically, based on measurement of a grey card. Software was developed to analyse images and to measure slice characteristics. Slices were automatically segmented from the background and their dimensions and shape were measured. Cells were segmented from cell walls using an algorithm designed to be tolerant of variations in slice reflectance and illumination uniformity. The position, area, brightness, elongation and orientation of each cell were measured. Statistical analysis of the cell size distributions provided information on the fineness of the cell structure. The spatial distribution of cell properties was also analysed to provide information on large-scale structure, and to relate this to conditions such as moulding. Examples are shown demonstrating use of the method to measure effects of processing conditions and ingredients.