Solid Rocket Motor X-ray Detection Film Defect Recognition Technology

Traditionally, the defect detection of solid-rocket engine in China is manipulated by technical personnel by means of X-ray crack-detection film. The automatic level for this manner is lower and also often makes mistakes caused by technical personnel's subjective factors. This paper tries to change this situation by doing the following research works:1. Make use of edge detection technique based on wavelet transform to find out the edges in the image and detect some no-clear defects derived from X-ray crack-detection film. The result comparisons between this approach with the traditional ones are given.2. Define some statistical gray scale parameters to a certain area in the image, these parameters are calculated and put into a BP artificial neural networks (ANN) as the input pattern to judge if there is possible defect existed in the area. In this part, this paper will put the emphasis on why and how to get these gray scale parameters, and demonstrate the result of judgment at last.3. Take advantage of various digital image process techniques to help technical personnel judging defects. In this part, this paper will mainly introduce aprototype system-REDRS(Rocket Engine Defect Recognition System),which is used to check the defects derived from the X-ray crack-detection film image of solid-rocket engine. There are many digital image process techniques are used to help the technical personnel to understand the image, this paper will summarize the theory and demonstrate the result.In the end, some further research propositions are given.