Application of X-ray digital radiography to online automated inspection of interior assembly structures of complex products

The paper proposes an application of X-ray digital radiography to online automated inspection and recognition of the interior assembly structures of complex products by means of the multiple views techniques. First, a vertical hybrid projection function (VHPF) is proposed as the recognition feature of a two-dimensional image. VHPF combines an integral projection function and a standard deviation function so that it can reflect the mean and the variance of the pixels in the vertical direction in an image. Secondly, by considering the different importance grades of objects inside the product and the independence of these objects along the circumference, the paper presents a hierarchical recognition method and uses a neural network system to speed up the computation process with parallel operations. Thirdly, using the whole-orientation features of one standard swatch and by extracting its maximal system of linear independence as the feature basis, the issue of blind areas for recognition is resolved. Based on this approach, the first domestic X-ray multi-view digital detection system has been developed and applied to the online detection of objects containing complicated assembly structures.