X-ray imaging is the most effective way to solve the correctness of fit of the internal components.Although the previous detection method

which is based on extracting and matching characteristics such as the shape of the connected regions in sample images

the aspect ratio or the area

achieves a better detection result.But by mechanical precision

assembly tolerances

parts dislocation and other factors

it is less robust.To solve this problem

by combining convolution neural network classification with CT technology

this dissertation first designed a deep Convolutional Neural Network model.Extracted features and trained the classifier using deep learning methods to classify the internal parts of a workpiece

outputting the coordinate frame and performing missing part detection.Then

uses the coordinates of the detected parts

we find standard workpiece views that conform to the current test-workpiece's angle based on CT projection sinusoidal properties.Perform assembly error detection

such as transposition or dislocation of the test workpiece's internal parts.Experiments show that our combined approach can identify missing and misaligned internal parts of a workpiece.The overall system is robust to factors such as overlapping among the internal parts of the workpiece.