Research On The Edge Detection Method Of Blood Vessel Wall Images

The processing of medical digital images has become one of the fastest growing areas in the world of medicine.It has been widely used in clinical diagnosis and treatment,which can enable clinicians to observe the lesions inside the human body more directly and clearly,and improve the diagnostic rate of diseases.It has become a crucial technology and method in the field of clinical medicine.The objective analysis and processing of medical images is an extremely important part of the imaging technology.In particular,smooth filtering and edge detection techniques involved in image processing technology has become the current research direction.Researchers can use some image processing techniques to process different types of medical images effectively to improving the use of information in medical images.It helps to achieve the interception of relevant features of lesions and further improve the accuracy of diagnosis and treatment.This thesis is based on the edge of the inner wall of a company in Wuhan City,and carries out the edge detection of OCT images of coronary artery vessels.After the laser probe collects the OCT image of the blood vessel wall,edge detection is performed to provide an intuitive image basis for follow-up surgery.Therefore,the detection must require real-time and precision.This standard requires that the complexity of the algorithm we choose cannot be high,but the accuracy of the detection method must be high.For this reason,this thesis first studied the theory of the existing image edge detection technology and chose to use a simple,stable and mature Canny algorithm for detection.Then,an improved algorithm is proposed for the shortcomings of the traditional Canny algorithm,followed by experimental comparison and analysis,and applied to the edge detection of blood vessel inner wall images finally.The following work is specifically completed.Firstly,this thesis objectively introduced the collection method and process of OCT images of coronary artery vessel wall,analyzed its characteristics,and studied the difficulties and requirements of its detection.For the needs and difficulties of the inspection project,the image is preprocessed to prepare for the next inspection.Secondly,the principle and disadvantages of the Canny algorithm with better comprehensive performance are analyzed in detail.For the Canny algorithm,the limitations of filter denoising and artificially setting the high and low threshold uncertainties,we decided to proceed from two aspects of filter denoising and threshold selection automation to improve the traditional Canny algorithm.Then,the method of filtering and denoising was analyzed.After comparison,the Top-hat method of mathematical morphology was selected to filter and denoise.As improved algorithm applied to the edge detection of intracoronary OCT images of coronary vessels,the detected edge details are more abundant.Finally,the automatic method of threshold selection was analyzed.After comparison,the largest inter-class variance method was used to select the threshold automatically.The thesis used improved algorithms for edge detection of coronary artery vessel wall OCT images.The detection effect has been further improved,and the smooth filtering capability and automation level have been effectively improved.The result of the detection is clear and precise,and it also meets the requirements of real-time performance and meets the needs of the project.