Rader image target pattern recognition based on Least Risk bayses classification

We define a L × L template when drawing eigenvalue in every graphic on the basis of binary data of Radar image, accumulate in statistics for each pixe and divide the area of this template, Setting threshold T. Setting eigenvalue 1 if the share of template black pixel is more than T. We use bayses classification to recognise the rader image targe, we choose the parameters on the basis of Least Risk rule. At last, we analysis the reasons for wrong classification, and state the Relevant factors.