Submillimeter Crack Detection Technology of Eggs Based on Improved Light Source

Abstract

Abstract The purpose of this paper is to provide an effective method for detecting submillimeter crack in eggs. The central idea is to enhance the crack information, because the submillimeter crack information is easy to be lost. Different from the pressure method used in previous literatures, this paper improves the light source for image acquisition. The laser light source is placed on the back of the egg, and the crack information is enhanced by the feature of high laser brightness. But the resulting problem is that stomata and eggshell edges are also noisier. To solve this problem, this paper comprehensively utilizes the characteristics of fixed threshold and adaptive threshold methods. The feature of fixed threshold is that it can select global threshold to eliminate the edge noise of eggshell with small pixels, but the stomatal noise is larger. The characteristic of adaptive threshold is that it can reduce the stomatal noise but keep the edge noise. According to the above analysis, this paper applies the two methods to obtain two result images at the same time, and performs and operates on the result image to filter the noise of egg stoma and eggshell edge, and determines whether there are cracks based on the standard of area. In the experiment, 100 eggs were tested, and the recognition of tested, cracked and comprehensive eggs reached 90.0%, 92.0% and 91.0% respectively. The innovation of this paper is to enhance the crack information in essence by improving the light source, instead of improving the recognition degree only by improving the algorithm. The comprehensive use of fixed threshold and adaptive threshold is also one of the innovations of this paper. Previous journals only improved or compared one threshold selection method, but did not make comprehensive use of their advantages.