Research And Implementation Of Vision-based Skin Lesion Plaque Measurement System

As the largest and most important organ of human body,skin is closely related to human health.With the continuous development of industrial science and technology,people’s pace of life is accelerating,environmental pollution,life anxiety and other problems are full of people’s life around,leading to the types and incidence of skin diseases in recent years are increasing;At present,most measurement methods for skin lesions and plaques are contact measurement or direct visual measurement.Such measurement methods are highly subjective,low measurement efficiency and easy to bring discomfort to patients in the measurement process.Therefore,it is of great significance to study the medical equipment that can realize the rapid measurement of skin lesions and plaques to assist practitioners in the evaluation of the efficacy of skin disease treatment.In this paper,the typical skin disease "vitiligo" is taken as the research object,and FPGA is used as the realization platform of image processing algorithm to research and design a vision-based skin lesion plaque measurement system.Firstly,this paper studies the realization of color image grayscale,verifies different realization methods,and finally uses weighted average method to realize color image grayscale.Then a variety of image denoising algorithms are proposed based on the possible noise of the images collected by the OV5640 camera.Finally,the median filtering algorithm is adopted as the image denoising algorithm through comparative analysis.Then,the traditional edge segmentation algorithm is analyzed through experiments,and an improved edge segmentation algorithm is proposed based on Sobel algorithm combined with nonmaximum suppression and median method to improve the accuracy of white spot edge segmentation.Then the threshold segmentation method is studied.Taking the typical OTSU algorithm as the realization object,an image segmentation algorithm combining improved edge segmentation and threshold segmentation is proposed finally,which improves the accuracy of white spot region segmentation.Finally to simplify the above algorithm,designed to be suitable for the FPGA implementation of the algorithm,in the design of a good video image acquisition system based on FPGA,the above algorithm implementation,realized with white spot region segmentation,and then in the FPGA system of white spot area in every frame image pixels,through a single pixel area,known finally achieved in the area of the white spot area measurement.The system uses the characteristics of FPGA parallel processing to realize the realtime segmentation of the white spot area and calculate the area of the area,overcoming the characteristics of portability and real-time.In today’s high incidence of skin diseases,it is of great significance to assist the diagnosis of practitioners and has a broad development prospect.