1. Introduction
With the rapid development of science and technology, complex surface parts have a wide application in machinery, electronics, medical science, etc., so the proportion of complex surface is increasing in manufacturing industry. It is an important task of manufacture to control the machin- ing quality of complex surface parts. Machining quality control means not only controlling external quality, such as machining precision, and surface quality, but also control- ling the internal quality of parts. There are various methods of inspecting the internal defects of complex surface parts, e.g. radiograph, eddy flow, ultrasonics, etc. Amongst these testing techniques, ultrasonic testing has an important role due to its strong penetration, good direction, high sensitivity, low cost, and being harmless to the human body and the material of the parts. In traditional ultrasonic non-destruc- tive testing (NDT), according to the ultrasonic echo, the testing personnel can only judge internal defects by his or her own experience. Because the traditional inspectionmethod is too subjective, the inspection reliability and efficiency are low. With the rapid development of the technologies of the computer, signal analysis and proces- sing, the digitalization, imaging, intelligence and the auto- matization of ultrasonic NDT has become a research focus. Particularly, the application of the ultrasonic imaging tech- nique and the automatical identification defects technique, can not only use an image to show the internal defects of parts, but can also use the modern signal analysis technique for quantitative and qualitative analysis and non-destructive evaluation of the defects. Thus it has an important signifi- cance for controlling product quantity, to research on an intelligent ultrasonic NDT system for complex surface parts.