VISUAL FARIS FULLY AUTOMATIC RADIOSCOPIC INSPECTION SYSTEM Visual Faris features the state of the art of imaging technology on automatic defects recognition for industrial X-ray applications. This innovative artificial vision system , totally designed and developed at Bosello High Technology imaging research laboratories, provides reliable and effective diagnostic tools for all applications that requires quality controls on light and heavy alloy castings. Visual Faris x-ray inspection consists of the processing of a series of images taken from a casting in positions that have been programmed for the inspection. As soon as the part of the casting to be inspected has been defined, a sample image is acquired for every position. These images are then used and enriched with data during the programming process, which consists of the setting of all the parameters and the necessary information for processing, analysis and interpretation of the detected defects. REGIONS OF INTEREST ROI identifies one or more areas within X-ray image where processing functions are executed and where specific rejection criterias and thresholds are applied. If different rejection thresholds have to applied to specific areas, different ROI can be created and associated with areas that have to comply with the same critical requirements. Rejection criteria are based on single defect features as well as on neighbourhood relationship in terms of inter-defects-distance and concentration. Custom rejection criteria can be implemented and integrated into Visual Faris functionalities. IMAGE ALIGNAMENT Visual Faris does not require a high mechanical precision concerning the casting’s position. It is able to correct possible misalignment between the image that has been acquired on-line and the one acquired during teach-in procedure. The alignment correction is carried out on the digital image and not by a mechanical repositioning of the casting. This considerably accelerates this operation up to almost real time speed. REMOVE ALIASING FEATURES The casting may have some “disturbing” details on its surface like writing, bar-codes, labels, etc.; these details on the image may disturb the analysis of it. The centre of the character “O”, for example, could be mistaken for porosity. For this problem Visual Faris offers a set of functions that allows to compensate side effects caused by this situations. These images show a typical situation where defects are located inside ID numbers. The first images shows the code "307763" and some porosity between "0" and "7". The second image show the result of removal algorithms, code numbers are totally removed while porosity is still visible. The third image shows porosity identified. PROCESSING Proprietary search engines have been developed by Bosello HT for defects identification and evaluation. Built in artificial defects are available for image analysis test and ADR programming debug. IMAGE QUALITY X-ray image acquisition is performed with specific imaging processes that enhance the information and reduce most of the effects that corrupt the image quality and contents. Noise reduction with integration technics. Field of view sensitivity equalization. X-ray absorption law compensation. Acquisition with multi energy technics. STATISTICS Visual Faris saves a variety of informations related with castings inspection : Production counters. Inspection results Defects features, geometrical , morphological, relational, etc... Images of good parts, bad parts, all parts.