Mathematical Analysis and Application on Mechanical Image of Hybrid Wavelet Transform Algorithm

To overcome the shortcomings such as significantly de-noising effect and easily losing the details of the image characteristics of the existing image de-noising methods, an image de-noising algorithm based on the hybrid wavelet transform was proposed. The algorithm integrated the advantages of wavelet de-noising retaining image details features and Wiener filter obtaining the optimal solution, and took the images processed by wavelet transform and Wiener filter as male and female of the initial population. The steps of the algorithm are as follows: mapping from image space to coding space, iterating to parents through selection, crossover and mutation operation until the offspring meeting the constraints was obtained, reducing the superior offspring to image space, gaining the approximate optimal solution. Theoretical analyses were made on the core of the algorithm, coding, crossover and mutation. The algorithm was applied to agricultural machinery parts image de-noising such as plough and disk harrow. The results showed that it had the advantages of high peak signal to noise ratio (PSNR), obvious edge characteristics, good vision effect, and so on. The result of the present work implied that the proposed algorithm is an effective and feasible exploration.