Studies On The Spectral Image Processing Technology Based On Gyrator Transform

In this dissertation, self-imaging, image filtering, noise reduction, andencryption of spectral images are investigated through gyrator transformation. Theresearch results are as following:1. Periodic images can be reproduced in the gyrator transform domain under propertransform angles, which means the gyrator transform Talbot effect can beobtained. Such Talbot effect can resist pollution to some extent, i.e. it hasself-cleaning capacity.2. Low filtering and band-stop filtering can eliminate the details of the gyratortransformed spectral images, and the original spectral images can be almostreproduced at high transform angles. However, high filtering and band-passfiltering can only remain the details of the transformed spectral images when thetransform angle is large.3. The amplitude of the gyrator transformed spectral image contains muchinformation of the original spectral image when the transformed angle is small,and the original spectral image can be almost reproduced from the amplitude.With the increment of the angle, the information of the original spectral image inthe amplitude decreases, that the original spectral image cannot be reproduced.The corresponding phase mainly contains the details of the original spectralimage, that the edge information of the original spectral image can be reproducedfrom the phase.4. Noises with hyperbolic type can be effectively reduced through gyratortransform. And, noises with hyperbolic information can be also reduced to someextent.5. Gyrator transform can not only encrypt gray images, but also encrypt colorimages. The encrypt keys can resist decryption strongly. The strength ofdecryption resistance can be strengthened by repeating the encryption. Suchencryption can be obtained by encrypting the spectral image directly, or byencrypting the amplitude and phase of the transformed spectral image in gyratortransform domain.