THE EXPERT SYSTEM FOR INDUSTRIAL RADIOGRAPH EVALUATION WITH IMAGE ENHANCEMENT
Bagus Tri Atmoyo, Djoko Marjanto, Supriyono
Sekolah Tinggi Teknologi Nuklir - Badan Tenaga Nuklir Nasional
Jl. Babarsari Kotak Pos 6101/YKBB Yogyakarta
The Radiograph Evaluation Expert System called REES has material data and radiograph data as input. Material data contains shape, welding technique and dimension (thickness and diameter) of the material. Radiograph data contains parameters of defect indication like shape, intensity, location, direction, distribution, continuity, and also thickness. The data are used as premise to select the rule to construct the conclusion. The conclusion contains information about defect appearance and kind of its, material acceptance, the cause and the solution. The radiograph data collecting process can be helped by image enhancement feature in the REES with digitization process of radiograph. Image enhancement feature contains brightness, contrast, saturation and exposure enhancement; histogram equalization; gray scale transformation; inverse intensity; convulsion and edge detection using Sobel or Roberts filters. The testing of REES used digital radiograph image courtesy of IIW as evaluation’s object. The image that had been actualizing with image enhancement feature show a defect indication with length shape, dark (low intensity) and located in root of the weld. Data processed by inference machine and found the conclusion informs that defect had been found. The defect is Incomplete Penetration. According to acceptance criteria of ASME VIII Division 2 – Alternative Rule, the material must be rejected and for remedial actions take apart the weld and do re-welding process to the material as the solution for this result.
Keyword: Expert System, Radiograph, Defect, Image Enhancement