A sub-pixel edge detection method and system

Abstract

The invention discloses a sub-pixel edge detection method and system. The method comprises the following steps: firstly, determining a gradient image according to an original image, determining a plurality of seed point pairs according to the gradient image, secondly, extracting edges by taking a pixel track corresponding to each seed point pair as a unit, and then determining a pixel edge contourof the original image according to the pixel track corresponding to each seed point pair; and finally, determining the sub-pixel edge contour of the original image according to the pixel edge contourof the original image by utilizing a spline interpolation method or a Gaussian curve fitting method or a Steger method. The contour tracking is realized by adopting a tracking idea, and compared withthe traditional pixel-by-pixel edge contour detection in a sliding window manner, the integrity of the contour is better kept.

Description

technical field [0001] The present invention relates to the technical field of sub-pixel edge detection, in particular to a sub-pixel edge detection method and system. Background technique [0002] The existing sub-pixel edge detection methods can be roughly divided into three categories according to their principles: curve fitting methods, interpolation methods, and moment methods. The fitting-based sub-pixel edge detection method, such as the method disclosed in patent CN107301636A, first uses the Canny edge detection algorithm to obtain pixel-level edge position information, and then improves the edge position accuracy to sub-pixel level by Gaussian fitting method. In addition, for example, CY Su et al. from Taiwan Normal University used the Canny algorithm as the sub-pixel contour detection method of the LED probe of the rough detector. The Canny operator needs to set the high and low thresholds and the Gaussian filter template size separately according to the specific s...

AI Technical Summary

Problems solved by technology

In addition, for example, CY Su et al. from Taiwan Normal University used the Canny algorithm as the sub-pixel contour detection method of the LED probe of the rough detector. The Canny operator needs to set the high and low thresholds and the Gaussian filter template size separately according to the specific scene, and has no parameters. Adaptive ability, and the detection result of the Canny operator is prone to contour breaks, resulting in incomplete target contours, and the amount of calculation in the fitting process has become a major bottleneck restricting its application

Sub-pixel edge detection based on interpolation methods such as Zhang Hongtao and Sun Qiucheng use B-spline functions and cubic spline functions as interpolation functions respectively. The method disclosed in patent CN106251327A uses bilinear interpolation functions. On the one hand, the selection of polynomial order It is a tricky problem: fitting high-order polynomials will greatly reduce the calculation speed, and using low-order polynomials often fails to meet the accuracy requirements

On the other hand, such methods are sensitive to image noise and often get poor results

The sub-pixel edge detection algorithm based on the moment method, such as the method disclosed in the patent CN104715491A, if the template effect is not considered in the gray moment edge detection algorithm, the calculation result will be inaccurate, and the template size needs to be selected according to the application scene to reduce the edge sub-pixel The deviation generated by the coordinate calculation makes the adaptability of the algorithm worse

The method disclosed in patent CN104899888A has relatively high computational complexity, making it unsuitable for industrial scenarios that require high detection speed

Patent CN104715487A discloses a sub-pixel edge detection method based on pseudo-Zernike moments. Although the method is not sensitive to noise, that is, it overcomes the influence of noise, but because it uses a pseudo-Zernike calculation method, the computational complexity of pseudo-Zernike moments is relatively high. Large, then this will affect the speed of calculation, and it is also not suitable for industrial scenarios that require high detection speed

Yang Bingbing et al. combined pseudo-Zernike moments with classic methods such as Sobel operator and Canny operator to seek a sub-pixel edge detection method that improves detection accuracy, which can accurately detect digital image edges, but because the detected image is only Considering the case of low noise, the detection accuracy needs to be improved for the case of high noise in the image

In addition, when the two edges are close to each other, the proposed algorithm will detect inaccurately, and even the detection error will occur.

Images