The patterns formation model of the hierarchical structure objects 
on the gray scale images in view of both geometrical features of object 
and semantic importance of levels and objects is offered. The approach 
to its realization using wavelet transform that has properties 
of frequency-spatial selectivity, allowing to detect information elements of object 
(for example, contour) with required detailed elaboration, is considered. 
The developed model enables to create flexible models of images processing, 
to avoid necessity of repeated comparison of the received information 
with a lot of models during objects classification, to receive the universal 
approach for the decision of a lot of applied tasks. 
Keywords: contour analysis, model, gray scale image, wavelet transform.