Modulation transfer function measurement method based on dynamic initial value fitting

Modulation transfer function1/4 MTF1/4 quantitatively characterizes the quality of the optical imaging system. On the basis of the ISO12233:2017 standard, a method for fitting super-sampled edge spread function (ESF) with three Fermi functions based on dynamic initial values is proposed, and the Canny algorithm with automatic threshold is used to detect the edge line. The proposed method dynamically adjusts the initial parameter values of the three Fermi functions obtained by the empirical formula, and uses the least squares method to fit the super-sampled ESF data. Compared with the ISO12233:2017 standard, the proposed method greatly reduces the influence of noise, which is mainly reflected in edge detection and the super-sampled ESF obtained from the projection of the image data. Simulation experiments are conducted on images with different blur and noise levels. The results show that compared with the method of empirical formula to get the initial fitting value, the proposed method is superior to the former in accuracy and scope of application. The experimental setup of MTF is established, and the same conclusion is obtained in actual measurement.