Sub-pixel edge detection method based on Gaussian fitting

Abstract

The invention discloses a sub-pixel edge detection method based on Gaussian fitting. In the method, a Canny edge detection algorithm is firstly used for obtaining pixel-level edge location information, and then the edge location accuracy is promoted to the sub-pixel level through a Gaussian fitting method. Experiments of synthetic and real images verify the accuracy and stability of the method provided by the invention, and comparison with other similar algorithms illustrates the advantages of the method. The method has broad application prospects in many computer vision application scenarios such as quality detection, remote sensing image processing and 3D reconstruction.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a sub-pixel edge detection algorithm based on Gaussian fitting. Background technique [0002] With the development of machine vision, vision systems equipped with CCD cameras are widely used in the fields of remote sensing, measurement, quality monitoring and three-dimensional reconstruction. Most of the applications in these fields need to analyze the edge in the image to obtain the information of the scene, so the accurate extraction of the edge information in the image plays a vital role in the realization of the function of the vision system. Traditional edge detection algorithms such as Canny and Sobel algorithms can only provide pixel-level accuracy, which means that the edge information between integer pixels is lost. In view of this situation, the sub-pixel edge detection algorithm came into being. At present, the sub-pixel edge detection algorithms can be main...

AI Technical Summary

Problems solved by technology

Usually this type of algorithm uses polynomial fitting method to realize the calculation of interpolation, but the selection of polynomial order is a thorny problem: using high-order polynomial fitting will greatly reduce the calculation speed, and using low-order polynomials often cannot achieve Accuracy requirements

These algorithms can achieve better results when dealing with noise-free synthetic images, but if the image contains noise, the algorithm may produce large errors

Images