Sub-pixel Edge Detection Based on Logistic Function Fitting

In machine vision size detection, it is often necessary to locate and extract the object edge, but the traditional edge detection algorithm can only locate the actual edge to a certain pixel when detecting the edge. Improving the camera resolution can improve the detection accuracy, which will increase the cost of machine vision system. in order to realize high precision machine vision measurement system, this paper propose a subpixel edge detection algorithm based on that fitting of traditional gradient direction and Logistic function multi-coordinate points to improve the measurement precision of machine vision. The experimental results demonstrate that the precision of the proposed algorithm is better than that of B-spline interpolation algorithm, and the proposed algorithm well satisfies the requirements for anti-noise, high precision and strong real-time performance in machine vision measurement.