Analysis Of OSAHS Early Pathological Images Based On Top-Hat Operator

Obstructive Sleep Apnea Hypopnea Syndrome(OSAHS)is a condition of the upper respiratory system which is cause of the upper airway is not smooth that the patients in the sleep process is Sleep apnea,hypopnea,hypoxia and other pathological behavior.This disease is mainly reflected in the upper respiratory tract cramped,bent and other phenomena.Currently,the primary examination of this disorder is gathering medicine image by an electronic laryngoscope,then through the doctor’s task is heavy and subjectivity.To address these problems,by implanting image processing methods in electronic laryngoscope,to calculate the relevant medical parameters.This paper introduces the basic principles of mathematical morphology and Top-hat algorithm,and its application in medical image.By comparing with some common operators and wavelet transforms,this paper presents an algorithm based on morphological gradient edge detection.When processing the OSAHS medical image,image acquisition,image enhancement,image edge detection and segmentation,image filling and area calculation are introduced,also introduces the processing method for the diagnosis results of electronic medical image.Finally,algorithm in this paper compared with Roberts,Prewitt,Log,Canny algorithm,image processing method to obtain a morphology based on Top-hat algorithm,the target edge image is more clear,complete,edge closure and SNR is higher,the electronic diagnostic results more accurate.This paper takes oral image,internal nasal passages image and throat at vocal image for example,after image filtering,using the appropriate structural elements of arget region of image size,shape,through the Top-Hat algorithm to suppress multiple superposition of the background,and then through the structural elements of various shapes to image edge detection,get the filter level as the weight to achieve the final edge of the synthesis,we can get a full,smooth,low noise edge,finally using the edge segmentation and calculation of the area,according to area comparison,electronic diagnosis.In this paper,the edge of the mouth,throat and other parts of the image are calculated,and the degree of closure is more than 98%,and the signal to noise ratio is 19 db,much higher than Roberts,Prewitt operator,etc.