Noise types and formation mechanism of machine vision

Surface defect detection based on machine vision is more and more widely used in the industrial field. However, the noise in the image is inevitable, and the quality of filtering and denoising has a decisive impact on the success rate of subsequent defect recognition. This paper summarizes the causes and types of noise in machine vision images. This paper introduces three kinds of noise in detail, including Gaussian noise, Poisson noise and salt and pepper noise. The generation mechanism of the above three kinds of noise is analyzed and displayed in the detection image of electric vehicle magnetic tile components, and the characteristics of the noise are compared in the form of three-dimensional diagram. Finally, the common denoising methods are introduced.