Precise subpixel edge localization based on partial area effect

With the development of science and technology, whether in the industrial field or in other fields, the requirements for the accuracy of image measurement are getting higher and higher. A sub-pixel is a virtual pixel defined between two physical pixels of an image acquisition sensor, and its data type is a floating-point number. If sub-pixel technology is used to increase the accuracy to 0.1 pixel, it is equivalent to increase the resolution of the image system analysis by a factor of 10. For the model shown, this paper uses accurate sub-pixel edge localization based on partial area effect, and uses Gaussian low-pass filtering for filtering to deal with edge glitches and edge shadows. At the same time, for other models, this paper uses Hough transform to fit separately to solve the problem of sub-pixels.