Autofocus algorithm based on Wavelet Packet Transform for infrared microscopy

摘要

A new autofocus algorithm based on Wavelet Packet Transform (WPT) was presented to find the best focus of infrared microscopy, which is applied to test the defocusing amounts of athermalized infrared optical system. According to the statistical analysis of the gradient of image, the focus function was obtained by weighting the decomposition coefficients of wavelet packet. Experimental results demonstrated the algorithm had high sensitivity and accuracy for the true focus position from a series of infrared images. Furthermore, the autofocus algorithm was compared with some well-know algorithm like the Sobel Gradient(SG), the Sum-Modified Laplacian(SML) and the Discrete Cosine Transform(DCT) method in order to prove its performance.