A new attempt to the detection of the impact craters and other morphologic structures on the Earth's surface was made using high-tech computational methods. For detection and allocation of such structures, the fast two-dimensional wavelet transformation and the fast nonlinear multiparameter regression analysis for the digital elevation (DEM) data subsets and 3D shaded images, which are drawn using these data should be applied in combination. The computing process for the solution of the indicated problem takes the form of a computational experiment and includes, in particular, the procedure of training on the known reliable impact structures, the procedure of testing of the constructed regression model and detection and identification of the required shapes on the Earth's surface with their subsequent contrast improvement. This technology was used for the detection of the big circle structures on the territory of North America and Central Siberia. In this case, the GTOPO 30 data were used. The results obtained seem to be promising. The known circle structures are effectively detected and clearly distinguished on the surface area under study. The quantitative information is represented in convenient form for subsequent processing and analysis.