The Research On Determination Technology Of Cotton Maturity Based On Image Processing

Automatic cotton fiber maturity determination deals with many research fields such as image processing, pattern recognition, computer vision. The traditional cotton fiber maturity determination method has many shortcomings. With the continuous development of computer image processing technology, the research of cotton fiber maturity automatically determination has made great progress. However, the research about automation cotton fiber maturity determination by using computer technology is still a considerable difficulty. For the application and research of the field is still relatively few at home and abroad, there are still many problems to be studied and solved. The focus of this research is to achieve computer automatic cotton fiber maturity determination by using the digital microscopic images of cotton fiber.Automatic cotton fiber maturity determination system consists of three steps. First, preprocess the microscopic fiber cross sectional images obtained before (including uneven illumination removal and cotton fiber image edge detection, etc). After that, extract a single fiber from the processed image, and then statistic geometric parameters of the cotton fiber by using the area fill method and chain code tracking. Finally, calculate the cotton fiber maturity by the statistic data. Among these steps, the most critical technologies are image preprocessing, image segmentation of cotton fiber, the adhesion cotton fiber separation, extraction of cotton fiber features. The main research content of this paper includes image enhancement of cotton fiber processing, separation of the adhesion cotton fiber in the cotton fiber image segmentation and extraction of cotton fiber features.Due to the limitations in the biopsy sample process on the production of cotton fiber, there is some negative impacts on the accuracy of the follow-up algorithm, such as background noise, cotton fiber shape changing, adhesion cotton fiber in the grayscale images sampled by microscope, etc. Therefore, removal of background noise interference, extraction of foreground part containing cotton fibers from the image accurately is the first major step in the cotton fiber maturity determination system. In order to find the foreground area containing cotton fibers, to accurately determine the cotton fiber is very important. In this paper, we propose an image segmentation algorithm through the study of the human visual mechanism model and the characteristics of cotton fibers. The experiments prove that the algorithm can part the foreground area containing cotton fibers from the image, which ensures the accuracy of the subsequent steps required for the input image.The accuracy of the cotton fiber segmentation has a direct impact on the cotton fiber maturity determination exact rate. In this paper, we comprehensive analysis of the limitations on the characteristics of commonly used algorithms in the field of image segmentation and in the cotton fiber image segmentation applications. According to the characteristics of the cotton fiber image, we apply difference of Gaussian offset filter to the segmentation of the cotton fiber. This algorithm can extract the cotton fiber image from the background accurately, also can suppress image noise effectively, and solve the inconspicuous edge of the cotton fiber problems, to ensure the accuracy of the subsequent algorithm.The cotton fiber in the foreground area segmented from the image background has adhesion and edge information of it is inconspicuous. In order to extract the final single cotton fiber, we still need to enhance the edge of cotton fiber, as well as adhesion-separation work. By studying edge detection, adhesion fiber separation and other related algorithms, this paper presents the use of the algorithm to enhance cotton fiber contour information based on Directed Graph, and adhesion separation algorithm for the separation of cotton fibers based on Euclidean distance. These algorithms are able to fill the contour defect, achieve the goal that accurate separation of the cotton fiber. By using seed filling and chain code combination algorithm counts the characteristics of the cotton fiber parameters, we calculate the cotton fiber maturity ratio, and ultimately come to the maturity of cotton fibers. Experiments show that the cotton fiber maturity ratio obtained by our method consistent with the actual cotton fiber maturity.