We reformulate the Poincare wavelet transform, proposed by Gorodnitskiy and Perel in Proc. DD'll [1 J, in the framework of li g ht cone coordinates. The rotation matrix turns to be dia g onal in li g ht cone coordinates.The interest to application of wavelets in quan tum field theory problems is caused by the existence of physical phenomena poorly described in terms of the plane wave decomposition of the wave func tion. This interest is also stimulated by the search for quantum field theory models free of ultra-violet divergences [2]. For instance, the registration of a single photon with the help of photographic plate results in local exposure of the plate, in contrast to theoretical expectation of global exposure for the photon treated as a plane wave.Let us briefly recall the application of wavelet transform to 2D image analysis, specially in cases where the behaviour of the analysed function in dif ferent directions is investigated at different scales.Let G be the affine group in JR d :Having calculated the wavelet coefficients (4) one can analyse the local behaviour of the function f at each point b E JR d at a given scale a in the direction ¢ E SO(d).The most known example is the analysis of im ages. In the case of d = 2 the group of rotations SO( 2) is an abelian group. Its two-dimensional rep resentation is given byThe Haar measure on SO( 2) is dp,(¢) = d¢ and the reconstruction formula is straightforwardwhere M (¢) is a unitary representation of the ro-where tation group SO(d) c G. Let r 2 1 � ( k ) 1 2 Cw = J lR 2 d kl'k'P < 00.(7)be the representation of the affine group (1) in JR d , constructed using a basic function 'I/J E L 2 ( JRd ) , which satisfies the admissibility condition [3] 1 r 2 Cw = 11'l/J112 J e I ('l/J IUI'I/J) I dp,da , b, ¢) < 00 , (3) with dP,L (a, b, ¢) being the left-invariant Haar mea sure on G. Here, for the sake of field theory applica tions, we used L 1 normaliz ation of the group repre sentation ( a I d ) , instead of the usual factor ()/2) of the L2 norm. The wavelet coefficients of a function To analyse anisotropic data in different directions the basic wavelet is chosen in the form stretched along one of the axis, like that shown in Fig. 1. Ro tating the basic wavelet in the range 0 :s; ¢ < 27r the coefficients (4) are obtained. Their depen dences on scale a may be different in different di rections. This is important for example in the anal ysis of anisotropic turbulence data [4]. We would like to have an analysing tool, similar to (6) , for the Poincare group, which includes hyperbolic ro tations of the (t, x ) plane. This rotations are rep resented by the matrix M(¢) = ( co� h¢ -sm h¢ -sinh¢) cosh¢ (8 )