Research On Edge Detection Algorithm Of Medical CT Image And FPGA Implementation

Medical imaging is widely used in clinical practice,including four major types:Computer X-ray Computed Tomography(CT),Magnetic Resonance Imaging(MRI),Positron Emission Tomography(PET)and Ultrasound Imaging(UI).These medical images contain rich information,and correctly detecting the contours of organs and the details inside the organs can help doctors make accurate diagnoses and treatments.The processing methods of medical images can be roughly divided into the following categories: medical image enhancement technology,medical image visualization processing technology,medical image edge detection,segmentation processing technology and noise reduction processing technology.This paper mainly studies the edge detection algorithm of medical CT image and its FPGA implementation.The algorithm also involves some image segmentation processing,noise reduction processing and visualization processing,providing effective segmentation and detection observation for images,improving the reliability and visibility of images,aiming to improve the quality of medical services.To this end,this paper has done the following work:1.In order to solve the problem that it is difficult to achieve superior performance in terms of visibility and edge connectivity at the same time for image edge detection,based on the concept that contours are important edge sets,a Canny edge detection algorithm with adaptive regional threshold extraction is proposed.The watershed algorithm is used to obtain the approximate contour of the image,and the precise edges detected by the Canny edge detection algorithm are used as references to correct the image contour frame.Then,the gradient statistics and variance calculation are applied to the segmented main frame contour by using Otsu’s method to provide adaptive regional threshold for Canny edge detection algorithm,and finally an image edge detection result with both visibility and edge connectivity is obtained.On this basis,technical adjustments are made in terms of contour correction using the proportion of common pixels of contours,and the proposed algorithm is applied to medical CT original images.The brain and lung organs with distinct contour characteristics are selected as test objects,so that the overall algorithm can not only detect organ contours but also detect small lesions on organs.The test results show that the proposed method has a greater degree of improvement in terms of objective data evaluation such as edge connectivity,quality factor and information entropy,and subjective visual evaluation compared with the comparison algorithm.2.In the FPGA implementation of Canny edge detection algorithm,a modified Canny edge detection algorithm is proposed to solve the problems of large calculation amount and difficulty in gradient direction circuit design.The three-level pipeline multiplication and addition circuit based on Kirsch operator is used to realize the gradient direction calculation,which replaces the direction calculation approximated by division and trigonometric function in the original Canny,which improves the hardware implementation speed to a certain extent and also improves the image processing effect.3.Combining the proposed algorithm and improved circuit design,data is input from the computer through UART serial port,and the result is output on the LCD display screen.A hardware platform based on FPGA chip is built to realize the hardware transplantation of medical CT image Canny edge detection algorithm with adaptive regional threshold extraction,verifying the feasibility of hardware implementation of the proposed algorithm.In summary,based on the difficulty of edge detection that it is hard to achieve superior performance in terms of adaptability and visibility as well as edge connectivity,the Canny edge detection algorithm with adaptive region threshold extraction is proposed and implemented on FPGA.After technical adjustment,it is applied to medical CT image processing.In the process of FPGA implementation,the difficulty of hardware implementation is analyzed and the gradient direction circuit is improved.Finally,the algorithm is hardware-implemented,which provides a powerful help for the diagnosis of diseases in a faster and more accurate way for medical CT image processing.