Industrial CT technology is a visual testing technology, the results of industrial CT scanning imaging can not only clearly and accurately reproduce the internal structure of the measured object, but also quantitatively give the geometric dimensions of the density composition and internal details of the substance, which has been widely used in industry and medicine.
However, due to the limitations of technical principles and the software and hardware conditions of the equipment, artifacts are inevitably generated in the reconstructed images, and the image uniformity deteriorates and the quality is greatly affected. There are many types of artifacts, such as ring artifacts, edge artifacts, metal artifacts, scattering artifacts, etcAmong them, in the low-energy ray industry, the beam hardening caused by the ray continuum makes many of the artifacts produced in the reconstructed image have no obvious discontinuity and no clear boundary, which makes it difficult to identify and easy to cause misjudgment by inspectors.
In the results of industrial CT scanning, the X-ray energy of beam hardening is a continuum, and the high-energy X-ray penetration ability is strong, and the low-energy X-ray penetration ability is poor, and the low-energy photons are more easily absorbed by the material than the high-energy photons. When penetrating the workpiece, the share of low-energy photons in the X-photon decreases, the share of high-energy photons increases, the average energy of the beam increases, and the peak of the energy spectrum shifts to the right (the energy spectrum hardens), which is called beam hardening. If the ray energy value is taken as the abscissa and the relative intensity is the ordinate, the transmitted ray spectrum shifts to the right relative to the peak value of the human emission spectrum. Beam hardening can cause inconsistent measurement data, and when scanning a cylindrical uniform workpiece, the CT value decreases with the decrease of radius, and the gray value in the edge area is higher than that in the middle area, resulting in an artifact similar to the shape of a "cup", also known as a cup artifact. The discontinuity of the goblet artifact is not obvious, there is no clear border, and it is difficult to identify. The goblet artifact differs from the ring and edge artifacts, both of which have sharper borders.
At present, the polynomial fitting correction method is commonly used in the research and interpretation algorithms for artifacts, which uses different correction functions for each detector channel at each projection angle. Because the fitting line attenuation coefficient is different from the actual line attenuation coefficient of the substance, the corrected projection value force is also deviated from the ideal value, and the standard deviation of the corrected image is increased, and the original noise is amplified. This is determined by the correction principle of the linearization correction method, which is an unavoidable error. There are also some new methods, compared with the polynomial fitting method, this method improves the correction algorithm in principle, for each index projection data, the method of multiplying the weight function and the current index projection data is used to harden the correction, the standard deviation of the corrected image is slightly reduced, the noise is not amplified, and the signal-to-noise ratio is increased by more than three times, and the correction effect is ideal.
The reasons for the cup artifact in X-ray CT images were analyzed, and the line attenuation coefficients under different thicknesses were obtained through the stepped model and the projection data of the measured object based on the hardening artifact correction method of the weight function, and the bearing number of the hardening model and the weight function of the correction model were given, and the weight function was determined. The results show that, compared with the most commonly used polynomial fitting method, the signal-to-noise ratio of grayscale images is increased by more than three times without amplifying noise after the correction of the cup-shaped artifacts of cylindrical workpieces, and the image boundary retention effect is better. Of course, at present, the weight function is aimed at the disk-shaped CT image, and different weight functions need to be selected to complete the correction algorithm for the image of other geometric shapes, which is also the focus of the next research and practice.
In addition to the cup artifact, there is also the ring artifact. In order to remove the annular artifact in industrial CT images, Professor Wang Jue et al. proposed a new correction method based on projected sinusoidal diagram to improve the quality of CT reconstruction images and the accuracy of subsequent processing and quantitative analysis. Firstly, the S-L filter is used to filter the original projection data to enhance the artifact information. Then, the filtered projection data is integrated into the line, and differential processing is used to further increase the difference between the artifact and the workpiece contour. Then, the differential projection data is interpolated and averaged according to the number of interpolations, and the position of the artifacts is automatically selected according to the normal distribution. Finally, the method of linear interpolation and linear extrapolation is combined to correct the projection data at the annular artifact. The results show that compared with the correction method based on polar coordinate transformation and wavelet-FFT filtering, the signal-to-noise ratio gain of the corrected grayscale image is 1688 dB, effectively eliminating ring artifacts while maintaining image edges and resolution well. Annular artifact correction is effective for intermittent detector failures or unstable performance.
The CT scanning modes commonly used in industrial CT include one generation of scanning and three generations of scanning, of which the third generation of scanning has higher efficiency, but it is easy to cause ring artifacts due to poor correction methods, so weakening or eliminating ring artifacts is one of the main contents that reflect the technical level of CT system manufacturers.