Study On Online Nondestructive Measurement Technology Of Artillery Barrel

The purpose of this thesis is to design a measurement system of the inner surface roughness and a defect recognition system of the inner surface based on the digital image processing technology, which are parts of the on-line nondestructive examination (NDE) of the inner surface quality of the gun barrel.The main content of the paper deals with the light choosing, image pre-processing, image segmentation, feature extraction, defect recognition and roughness measurement. On the basis of the researching on the existed theories and techniques, the paper chooses and designs the optimum algorithm be used in the researching systems.Through the comparison of the results that come out of the artificial recognition and the researching systems, the paper puts forward a defect recognition system, whose accuracy rate is above 96% and is able to identify four types of inner surface flaws such as defect point, scratch, bruise and rust spot. The inner surface roughness measurement system is able to judge 3 grade of roughness:Ra1.6, Ra0.8, Ra0.4.