Study On Enhancement And Color Correction Method Of Endoscopic Images

The endoscope is an optical device,and can detect the image information in the object,the medical endoscope can through the natural orifice on body or the surgical incision into the body,and show the internal structure of the human body for doctors that X ray can not display,the endoscope become an essential device in check the stomach,esophagus and small intestine.by using the endoscope,doctors can establish patients information and improve the accuracy in diagnosis process effectively.Photodynamic therapy guided by the endoscope can more clearly show the lesions for doctors,and really reflect the surrounding structures of the lesions,that make it has been widely used in recent years.Due to limitations of light condition by the endoscope and the complexity internal structure of the human body,the images take by endoscope are not very well,because of inadequate exposure make large area of dark in the image,doctor can’t observe the operation space clearly,and it will be the main cause of error diagnosis or even failure of the operation.In addition,in the process of heat treatment,the treatment of light will produce a mistake of color in the images,in will take great influence of observe the patient’s condition in the process of operation,and postoperative evaluation of lesions.So,it’s an urgent problem to be solved in photodynamic therapy by enhanced display and color correction of the endoscope images.(1)this paper briefly introduces the technical problems in the photodynamic therapy guided by the endoscope,summarizes the development process of the endoscope,resolve the technical concepts of photodynamic tumor targeting therapy,and introduces the advantages and disadvantages of traditional image enhancement.(2)A histogram equalization enhancement based color correction method is proposed.compared with each pixel value in RGB channel and minimum RGB maximum pixel value,and count the number,get the largest proportion of RGB channel,and use grayscale diminishing respectively compared with RGB channels in each pixel values and counting,get the color correction coefficient,Finally,according to the largest proportion of channel as themain channel,correct color deviation images by the color correction coefficient.(3)A multi-scale Retinex algorithm with color protection is improved.Taking into the color of red and yellow are more then blue and cyan in the endoscope images.it changed the concept that average distribution of the RGB channel in luminance space by redefine the "luminance" information,combined with the multi-scale algorithm and the gain/offset algorithm,in the end enhance the original image.