Agricultural image de-noising algorithm based on hybrid wavelet transform

The current image de-noising methods cannot remove the noise effectively, and they have the disadvantage of losing minutiae easily. A new de-noising method based on Hybrid Wavelet Transform was proposed in this study. Wavelet de-noising has the advantage of keeping the image’s detail information, and Wiener Filter can obtain an optimal solution. This algorithm synthesized the advantages of Wavelet de-noising and Wiener Filter. Firstly, the image de-noised by Wavelet was used as male parent of the Hybrid Wavelet Transform’s initial population, and image de-noised by Wiener Filter as female parent. Then, the individuals with fitness function of maximum between-cluster variance were evaluated. Through the hybrid and mutate operation, the gene recombination was realized, and then the superior gene of the two images de-noised was extracted by Wavelet and Wiener Filter. Finally, with the finite order hereditary algebra, an offspring image was obtained which has both advantages of male parent and female parent. The performance of this algorithm was tested by the red jujube images and wheat images. The results showed that images of red jujube and wheat de-noised by the proposed method had a higher PSNR (178.44 and 183.24) than those processed by conventional methods such as neighborhood average (176.76 and 175.16), median filter (174.79 and 173.13), Wiener filter (172.75 and 173.48) and Gauss filter (167.50 and 165.60) etc. The experimental results showed that the Hybrid Wavelet Transform de-noising method used on agricultural image had the advantage of high signal-to-noise ratio, and good visual effect. Therefore, the method proposed is effective and practicable.