Applied Research Of Auto-focus Technology In Tongue Image Acquisition System

With the rapid development of digital imaging systems, optimization of the performance, and the technology's progress of the corresponding image information acquisition, display, storage, People have become increasingly demanding the quality of images. Especially in the medical field, the image quality directly affects the level of medical treatment. The tongue inspection is one of the most important approaches of observation in Traditional Chinese Medicine Modernization four diagnostic. Its objective study is an important component of TCM. However, due to the image in the collection process will be subject to various conditions, as well as the complexity of the tongue's own internal structure and individual characteristics of the diversity of tongue, leading to the quality of the images obtained is not good enough, and seriously affected the subsequent analysis of the tongue.There is need to propose new subject about the evaluation of image quality and automatic focus on performance. The research of auto-focusing technology causes related attention. The research will further enhances the image measurement accuracy, measurement speed and measurement of the degree of automation, which can reduce labor intensity, and so has important practical significance.Much work of predecessors is summarized, and the development of the auto-focusing technique at home and abroad is introduced. Then the pretreatment of the image-related as well as the mechanism why digital image is focused in optics imaging system is also analyzed. Then, the three key modules in auto-focusing are analyzed. The modules are:the focus window selection, clarity evaluation function and the top searching. In these modules, the focus window selection is that which parts of the image need to be focused evaluation._Image quality assessment mainly refers to that in which ways to judge the quality of the images. The module of the top search means that how to access high quality pictures quickly and accurately. Firstly, in the focus window selection, by studying the unique characteristics of the tongue in the experiment, we can use the huge difference about the tongue's and around the tongue's RGB values, and select the profile as the auto focus initial window by using grayscale enhancement function and GVF-Snake model. Then, according to the relationship between upper-lower diameter and left-right diameter, adjust the direction of the lens movement both from the neutral and angle of horizontal intake. In this way, the right focus window is quickly found.Secondly, in image quality evaluation, this article discusses the image sharpness function and the nature about focusing evaluation function. In particular, the peak characteristics, of the image sharpness function were studied in detail. According to the features that H values and V values are concentrated in a smaller scale when the pixels of tongue are in the HSV format. From gradient computation method based on the line from the cross, an improved clarity evaluation method is put forward. It obtains a more accurate quality evaluation.Thirdly, the module of top search Is also the focus of focusing operation, which directly affects the image quality and speed of acquisition. In this paper, the improved climbing algorithm is adopted, by comparing and mapping the average evaluation value of N points, then reduce the N value and mobile steps when lens close to the focus position in focusing process. This method effectively overcomes the shortcomings that the measurement of definition function access to the local limit area.In this paper, finally, the discussion of the above modules is all verified by experiment and a tongue image acquisition system is design. The results show that three modules all achieved desired results, and as well as that the system can more quickly and accurately obtain a clear picture of needs.