In this study, the object recognition problem was solved using support plane method. The modelled SAR images were used as features vectors in the recognition algorithm. Radar signal backscattering of objects in different observing poses is presented in SAR images. For real time simulation, we used simple mixture model of Lambertian–specular reflectivity. To this end, an algorithm of ray tracing is extended for simulating SAR images of 3D man-made models. The suggested algorithm of support plane is very effective in objects recognition using SAR images and RCS diagrams.
In the paper we have proposed recognition of object by RCS diagrams method. For modeling the scattering field of 3D objects on underlying surface we had use widely known FDTD method. We have used for distance function in developing method conjugation indices with so-called support plane, is formed within feature vectors of recognition class. We have given the results of recognition experiments with three different methods: support vector method, correlation method with the average class vector and a new support plane method.