The Research On Assistant Estimate System For X-ray Detect Of Nuclear Power Station

Non-destructives test is not only to detect whether there is lacuna in welding line but also to estimate material in modem times. It includes measuring the number of lacunas and estimating the quality of welding with lacunas. Non-destructives estimate is more advanced than non-destructives test. In traditional way, people observe film by themselves with light, magnifier, measuring for bucking and recording tab etc. In this way, it is easily to be affected by devices, environment and people's spirit. It limits the improving of work efficiency. Manual-estimate couldn't satisfy the development of manufacture.In recent years, as the developing of technology in digit image process, digit film and estimate intelligence has became the future heading of the non-destructives test. The intention of the paper is to design a system to help person to estimate quality of weld with compute. The system will use a series of technology in digital image process to realize image enhance, sharpening and lacunas identify, then system differentiate the lacuna and sentence the degree of weld with the expert's knowledge and experience which had saved in system before.In this paper, the assistant system has to separate to two parts to introduce. The first one is digital image process. This part analysis the math model of digital film image and uses the modified method based on wavelet threshold for denoising which has better PSNR contrasting with traditional methods. It also adopts morphologic to pick up the edge of the image. According to analysis the feature of the lacunas, it selects some characteristic parameter to reflect the feature of the lacuna accurately. Another part is expert system. It has to use the fuzzy expert system because of the uncertainty of the knowledge and the experiences. The system has summarized much knowledge to build a repository which based on national standard, books and expert's experience. It also chooses the easiest way to express rules. It has a fuzzy expression and replacement algorithm by reason of fuzzy expert system. In order to save energy and never to find aim blindness, the system uses forward direction and reverses direction consequence. According to the matching instance of the fact, it adopts multilayer multidimensional reasoning and solves the conflict with preferential matching degree.