Volume measurement method based on industrial CT scanning technology

Abstract

The invention provides a volume measurement method based on industrial CT scanning technologies. The method comprises the following steps: 1) edge extraction: extracting sub-pixel edges of two-dimensional images by using Zernike moments; 2) polynomial fitting: performing polynomial least square fitting on edge points, and transforming fitted point coordinates into polar coordinates; 3) interlayer matching: performing equal phase-angle-interval sampling on fitting function curves in step 2), and establishing a matching relationship between contour points having a same phase angle; 4) interlayer interpolation: acquiring the contour points in an intermediate layer by adopting the interlayer interpolation algorithm; 5) top prediction: estimating a top point of a workpiece by adopting the gray extrapolation method; 6) volume measurement: calculating the volume of the workpiece by using the truncated cone method according to three-dimensional fitting contour data. Based on the Zernike moments and the least square fitting algorithm, the method achieves area and volume measurement precisions which are 5 times or even more than 10 times higher than previous algorithms, and a top absence problem caused by volume effects or incomplete scanning in actual measurement processes is solved.

Description

technical field [0001] The invention relates to an image volume measurement method, in particular to a volume measurement method based on CT scanning images. Background technique [0002] Industrial CT measurement technology is an important research content for the development of industrial CT detection technology from qualitative detection to quantitative measurement. Two-dimensional measurement based on CT images has a good foundation, but there are still many problems in three-dimensional measurement. Three-dimensional measurement is to extract useful information from CT images, and quickly and accurately obtain spatial three-dimensional data such as the volume and surface area of ​​the object to be measured. The current CT image measurement is still at the stage of human-computer interaction to select the region of interest, the measurement results are greatly affected by the operator, the accuracy and speed are not ideal, and it is powerless to measure the three-dimensi...

AI Technical Summary

Problems solved by technology

[0005] In terms of three-dimensional measurement, the researchers have done the following work: the expansion and erosion processing in mathematical morphology is used for the contours of two adjacent layers to generate the contour of the middle layer, which can better maintain the contour details and better Show the change trend of the contour between layers, but the accuracy is limited to the pixel level; the interpolation method based on gray level directly uses the weighted average of the gray value of the adjacent layer as the gray value of the middle layer, which is simple and easy to implement, but it will Boundary blurring occurs, boundary positioning is inaccurate, and measurement accuracy is poor; the basic idea of ​​the shape-based interpolation algorithm is to extract the contour of the workpiece to obtain a binary image, and then perform distance transformation on the points inside and outside the outline, then interpolate the distance image, and finally threshold the It is transformed into a binary image to realize the uniform transition change of the interlayer profile. The measurement accuracy of this method is limited to the pixel level; the 3D Facet model method uses a polynomial to fit the Facet model, and obtains the second-order zero-crossing point according to the 3D direction derivative information to obtain the workpiece. The three-dimensional edge points, and then calculate the volume of the workpiece according to the polyhedron method. This method has high precision, but the calculation amount is large, and it is easy to produce discontinuous contours. In addition, some scholars have proposed an interlayer processing algorithm based on layer thickness geometric analysis to achieve high precision. Three-dimensional display of geomedical images, the effect is better, but the requirements for the similarity of the contour of the image to be inserted are high, and the speed is slow

Images