The application discloses an evaluation method for precision of a sub-pixel level
edge detection algorithm, comprising the following steps: S1, simulating an edge as a straight line or a curve equation, and recording the straight line or the curve equation as an edge line A; presetting a minimum value G1 and a maximum value G2 of gray scales on two sides of the edge line A; establishing an image coordinate
system, and recording a distance from a center of an i-th row and j-th column pixel to the edge line A as d ij ; S2, taking the d ij as an independent variable, taking a gray scale value of each pixel point as a dependent variable, constructing a
function model, obtaining the gray scale values of different pixel points in the image, and generating a
standard test image; the gray scale values of the different pixel points are between the minimum value G1 and the maximum value G2; S3, detecting the
standard test image by using the sub-pixel level
edge detection algorithm to be evaluated, and obtaining detection values of each point on the edge line A in a real test image coordinate
system; and S4, calculating deviations of the detection values from the edge line A in the
standard test image, and taking the deviations as the precision of the sub-pixel level
edge detection algorithm to be evaluated. The method can construct an edge conforming to a real scene, and is more practical.