Sub-pixel Edge Detection Based on Curve Fitting

Aimed at the problem that it s difficult to improve the identify precision of the linear CCD scan image, a novel fast sub-pixel edge detection method for image measurement is proposed. According to the step jump characteristic of the image edge gray degree and grads, it can determine the pixel boundary of the CCD image through edge automatic detecting algorithm. Based on this, we can use the curve fitting method to make sub-pixel subdivision of the edge position of the image, and carry out curve fitting of the edge signal. Finally, use the least square method to compare the second order curve which is obtained by fitting with the threshold level, and then use it to get the formula. We can use the formula to work out the accurate position of the edge point, thereby achieve the sub-pixel edge positioning precision. The experimental results and analysis show that, compared with the threshold level comparison method, the straight line fitting method has more advantages, such as high repeated precision, good stability and so on. It can restrain the influence of random noise effectively, so it can detect the edge position of one-dimensional image effectively.