Binarization for Cross-section Imaging of M. Longissimus Using Dynamic Thresholding Method

Abstract

We considered the image acquisition system and algorithm of dynamic thresholding method for automatically obtaining the appropriate binary image of beef rib-eye area. Our proposed image acquisition system, in which the physical object was covered with non-reflecting glass plate and cooking oil was filled in the space between them, was able to exclude irregular reflection spots. The dynamic thresholding method consisted of three steps. First, the threshold value of a split region was determined by ordinary thresholding method. Second, the threshold value of each pixel was estimated with linear function of threshold values at vertexes of the target region. Finally, binarization was carried out for each pixel based on its threshold value. This algorithm is described herein. We were able to obtain the appropriate binary image using the proposed image acquisition system and the dynamic thresholding method.