Application of tree classifier in the image diagnosis of the adhesive quality about the multi-layered structure

When the adhesive quality in multi-layered structure (steel shell, under layer and insulation layer)which is used in the solid rocket motor is diagnosed,the resonance between the echoes is used to obtain the signal characteristics, i.e. delay time. Extracting the signal characteristics and recognizing the adhesive state is a pattern recognition process. The dichotomous branch tree classifier is designed to accomplish the pattern classification .The high quality diagnosis image is obtained.