Workpiece Anufacturing Error Detection Method Based On 2D CT Image-3D Design Model

When using a two-dimensional CT image and a design model to compare the workpiece manufacturing error detection,we must know the position of CT image in the design model.To determine the position,we often need to find a reference position that matches the design model on the workpiece.In order to obtain high-precision detection results,we usually need to finish machining the reference of the workpiece,and if the workpiece is detected to be out of conformity later in this way,all the previous achievements will be wasted,and the processing time and cost will be wasted.In order to solve this problem,this thesis proposes a method of workpiece manufacturing error detection,which does not need to determine the workpiece reference position,and automatically searches the design model of CT image matching through 2D-3D registration.The implementation of this method mainly includes the following contents:(1)Pretreatment stage.It includes filtering enhancement,edge extraction,point cloud slicing and so on.Firstly,the CT image is enhanced by bilateral filtering and linear transformation,and then the Canny/Devernay sub-pixel edge extraction algorithm is used to extract the edge points to obtain the accurate edge information of the CT image;the point cloud model is transformed from the Solid Works design model,and the point cloud slicing algorithm is used for equidistant slicing.(2)CT image matching stage.A matching method based on the number of contour and Hu moment is proposed.The CT image after edge extraction and all the point cloud slices are matched by similarity,and the slice with the highest similarity is selected as the initial design model to match the CT image.(3)CT image registration stage.In order to improve the accuracy of the design model matching with the CT image,more precise slices(reducing the slice thickness)are performed near the slice position obtained in(2);then these slices are registered with the set of edge points of CT image,and the slices with the minimum mean square error after registration are selected as the final and accurate design model.In order to speed up the registration speed of iterative closest point(ICP)algorithm,singular value decomposition(SVD)algorithm is used to perform the operation,and k-d tree is used to search the nearest point.(4)Error analysis stage.According to the registration results of CT image and the final design model,the manufacturing error of the workpiece is displayed and analyzed.The error is directly displayed through the color error nephogram of CT image,and the error is analyzed quantitatively.In order to verify the feasibility and accuracy of the method,the standard workpiece with known size is used for detection.The results show that the detection error is concentrated between 0-0.35 pixels,which has high detection accuracy.In order to verify the applicability of the method,two actual workpieces are detected,and the change of mean square error and the distribution of error points are described and analyzed in detail.The results show that this method can detect the manufacturing error of the actual workpiece,and has high practical value.