Research Of Morphology And Watershed In Chalky Rice Detection

It is of great significance to perform identification of chalky rice in many fields, such as food processing, grading and food supply import and export. This subject was focused on study of the basis for grading chalky rice. The percentage of chalky area is an important basis for grading chalky rice.Firstly, studies on chalky rice identification are introduced, and 5 identification methods are presented. Although these methods differ from each other, statistical information such as gray value is taken as the identification threshold to identify chalky rice. Different variety and light condition make the threshold methodology lack of adaptability and accuracy, and identification accuracy of chalky rice was affected. It is the object for many identification methodologies to improve chalky ice identification accuracy. Morphological watershed method, which is to divide up according to the changes in gray value segmentation edges, not only has position information but also calculates the percentage area of chalky area accurately. Compared with other identification methods, it has certain advantages.Secondly, more detailed theoretical description for application of mathematical morphology in image processing field, including definition and property of binary morphology and gray morphology. Then, the principle and application of morphological watershed algorithm in image processing is discussed, as well as the implementation process of this algorithm. Simultaneously, several techniques, including gradient segmentation, distance transform and tag control, to conquer over-segmentation are presented in this article, considering that the problem of over-segmentation exists in the process of image segmentation using watershed algorithm.Finally, application of morphological watershed algorithm in chalky rice identification is put forward according to the physical characteristic that chalky opaque exists in chalky area of chalky rice, and feasibility and veracity is proved by experiments. The processes of identification and calculation are presented, simultaneously, existed methods are compared with.It has been proven by experimental images and data that the performance of edge detection in this method is more excellent than that in traditional classical detection algorithm and that it has more exact edge detection accuracy. Vast simulation on chalky rice images has proven that identification rate of morphological watershed method is higher than that of fractal dimension detection algorithm, with that this method includes space distribution information of chalky area which makes it more accurate and visualizable to identify chalky rice.