Research On Image Processing Technology Of High-Definition Medical Electronic Endoscope

Although considerable progress has been achieved by the sharpness and contrast of conventional medical electronic endoscope system and the obvious change in shape or color of lesions can be find by the system, the tiny lesions of digestive tract, the flat early canceration and the hyperplasia is too difficult to diagnosis, which often leads to missed diagnosis. Therefore, High-Definition medical electronic endoscope system and relevant advanced image processing technology arise at the historic moment. At present, High-Definition medical electronic endoscope system and image processing system (image processing workstation CCU) in China are mainly relying on imports, and the related independent research and development products haven’t appeared yet.The purpose of this paper is to study the implementation of the image processing algorithms of the High-Definition medical electronic endoscope, the research content includes: the image data acquisition, the image reconstruction (demosaicking), the implementation of the typical image processing algorithms (color correction, white balance, outline stressed, metering mode, contrast enhancement, etc.), the principle of dynamic band imaging technology and the evaluation of the contrast effect between capillary and surrounding tissue with ten groups of commonly used wavelength of FICE. The main work includes the following five aspects:(1) the speed of image display has been improved by using the double-thread method, one is used for image acquisition and extraction of efficient image data, the other is used for image processing, display and storage;(2) a series of image processing algorithms have been realized in order to satisfy the needs of the practical application of medical endoscope, which include the color correction, contrast enhancement, white balance, outline stressed, metering mode, etc.;(3)a color correction algorithm of the endoscope system which is four-neighbourhood polynomial regression method based on the YUV color space has been proposed to avoid the slant lubricious phenomenon, and the correction effect has been compared and analysed;(4) the principle and concrete realization of FICE (Fuji Intelligent Chromo Endoscopy) technology has been elaborated based on the comparison of advantages and disadvantages of existing multispectral imaging technology;(5) the objective evaluation algorithm of comparing contrast effect between capillary and surrounding tissue with ten groups of commonly used wavelength of FICE has been proposed, which is the improved Otsu algorithm based on brightness contrast.The performance index of the high definition medical electronic endoscope system based on CMOS image sensor has reached the international advanced level, the effect of image processing functions has met the requirements of clinical practice, so it has practical application value. The frame rate can reach up to26fps on the image resolution of1280×800. FICE technology has been realized in the endoscope system, and the ten groups of commonly used wavelength of FICE have been set in advance, which can be easily switched. The objective evaluation algorithm of comparing contrast effect between capillary and surrounding tissue with ten groups of commonly used wavelength of FICE has been proposed for the first time, which provides a strong theoretical basis for clinical medical workers to choose appropriate wavelength combination of FICE.