Page 95 - Phase Space Optics Fundamentals and Applications
P. 95

76   Chapter Three


               which, in particular, yields the energy conservation law

                            ∞               ∞
                                    2           U         2
                              | f (r i ) | dr i =  |R [ f (r i )](r o )| dr o  (3.44)
                           −∞              −∞
                 We also remind (see Sec. 1.6) that the Wigner distribution is rotated
               in phase space under the RCT
                                        U
                            W f (r, p) = W R [ f ] (Xr + Yp, −Yr + Xp)  (3.45)
               Moreover its projection corresponds to the squared modulus of the ap-
               propriated RCT, which can be registered experimentally. These prop-
               erties are crucial for phase-space tomography, which permits one to
               reconstruct the Wigner distribution from its projections. The details
               of this method for the case of the fractional FT are clarified in Chap. 4.


               3.4.2 Similarity to the Fractional
                      Fourier Transform
               It has been shown 36,37  that any unitary matrix U s is similar to one
               U f associated with the fractional FT. Indeed, the unitary matrix has
               unimodular eigenvalues and can be diagonalized. The diagonal uni-
               tary matrix corresponds to the fractional FT, Eq. (3.15). Moreover, the
               matrix that diagonalizes the matrix is also unitary, and therefore we
               can write
                                   U s = UU f (  x ,   y )U −1      (3.46)

               where   x and   y and the matrix U are defined from the eigenvalues
               and eigenvectors of U s , correspondingly. Then we can conclude that
               any phase-space rotator is similar to the fractional FT. For example, the
               signal rotator and gyrator are similar to the antisymmetric fractional
               FT because


                            U r ( ) = U g  U f ( , −  ) U g −
                                        4                 4


                            U g ( ) = U r −  U f ( , −  ) U r       (3.47)
                                         4               4
               Note that due to the symmetry of the phase-space rotator matrices,
               such as U g (  ±  /4) =−U g (± /4), there exist various similarity
               relations (see Sec. 1.6.2). For example, we can also write for the signal
               rotator


                         U r (± ) = U g ±  U f ( , −  ) U g ∓
                                       4                 4


                               = U g   ±    U f ( , −  ) U g   ∓    (3.48)
                                         4                   4
   90   91   92   93   94   95   96   97   98   99   100