Applied Research On Badminton Feather Image Processing Methods

As society and economy develop, living standards improve, more people are concerned about their health.And badminton sport is popular in China which is a big country of badminton production and consumption, the prospect of badminton market is broad. To a large extent, badminton quality depends on feather pieces quality. Usually a high quality badminton is implanted 16 feather pieces whose curvature and camber is similar, but there are many stained feathers, such as deletion, insects bite, wrinkles and other defects, which affect the quality of badminton, thus it is nessary to detect these stains, this work is done by hand before, there is a labor-intensive, low productivity and high cost, now gradually using machine vision technology to do it, reducing production costs, improve production efficiency.The paper takes badminton feather imge as the object of this research, analyzing the hardware and software in the badminton feaeher detection system, as well as studying the poor quality of uneven illumination image enhancement processing and image denoising algorithms, achieving automatic detection of badminton feather. It has a important theoretical and practical economic value.This paper mainly has the following innovation points for the research of badminton feather image processing:1. This paper study a complete hardware and software of the badminton feather automatic detection system, analysing the lens, illumination mode, the camera image acquisition and other factors on the effect of image acquisition. It lays a solid foundation for the study of badminton feather image acquisition and image processing algorithms.2. Proposed an adaptive image enhancement method Retinex theory. After processing, it improves the quality of the image, the texture information of image are more clear. Experimental results show that because of the influence of badminton feather’s own characteristics and light environment, these make uneven illumination images. Through using this method, paper can achieve bettet image enhancement results compare with other common image enhancement methods.3. Using the mean shift method of image filtering denoising to process badminton feather image, it obtains better subjective and objective evaluation, then achieves a good effect of the extraction of feather quill.4. This paper proposes dual CV level set method to segment wrinkle defect of feather image, and get good results, achieves the intended purpose of the experiment.This article is based on Matlab and VC6.0 platform accomplishing algorithms and methods above, and take a simulation test, and achieves good results.