Research On Defect Detection Of Micro-CT Detector And Artifact Removal In Imaging

Due to its high spatial resolution Micro-CT(micro-computed tomography)is widely used in scientific research fields such as model animal experiments and material microstructure analysis.To achieve high spatial resolution,multi-energy microfocus X-ray source and scintillator flat panel detector are applied in micro-CT systems.Due to the small size of the high-precision detector pixels and the difficulty in processing the scintillator and the back-end circuit,some pixels respond to X-ray nonuniformly compared with other pixels.Such pixels are regarded as defective pixels.Due to the attenuation of multi-energy X-rays after passing through the object,the energy spectrum received by different pixels of the detector is not the same.The difference is related to the complexity of the scanned object,which further increases the complexity of the defective pixel response.The effect of defective pixel errors is accumulated in the reconstruction algorithm,and obvious ring artifacts are generated in the CT image,covering the structure of the object and degrading the imaging quality.In order to improve the imaging quality of the micro-CT,research on the correction of ring artifacts is carried out as follow:Firstly,the position of defective pixel is detected by measuring the response characteristics,and a correction method is proposed by exploring the correlation between the defective and normal pixels.Specifically,aluminum sheets with different thicknesses are used to control the X-ray intensity received detector,measuring the response curve of each pixel.Then according to the difference between the response of the defective pixel and the normal pixel,the gray morphology method is proposed to accurately detect the defective pixel position;According to the characteristics of its response curve,the defective pixels are divided into two categories: defective pixels with truncation error and defective pixels with nonlinear response.The truncation error repairing method and the nonlinear response correction method based on DCT-PLS are designed respectively to repair the two types of defective pixel responses.The experimental results show that the error response of the defective pixels can be repaired and the ring artifacts can be corrected by the proposed method.Considering that the method based on pixel response curve correction requires high stability of the output X-ray intensity of X-ray source,and the experimental process is complicated,this paper proposes a blind ring artifacts reduction method based on the projection image of the object.Since the defective pixels appear as vertical stripes in the sinogram,a horizontal curve is derived by summing the pixel values along vertical direction,thus the abrupt segments related to the defective stripes are enhanced notably,and a proportion coefficient based on the second derivative of the curve is taken as the indicator for the position and the severity of the defective pixels.Then,the detected defective pixels in the sinogram are relocated in the projections,an improved 3D block matching filtering(BM3D)algorithm is applied to restore the defective pixels in corresponding projection images.In the end,the tomographic images without ring artifacts are reconstructed from the corrected projections.In the experiment,a small piece of the motherwort's rhizome and a part of a mouse's lung are imaged by micro-CT,and the result shows that,compared with the other four state-of-art methods,the proposed method can effectively reduce the ring artifacts in reconstructed images,and has less impact on spatial resolution and contrast at the same time.In order to explore the ring artifact correction method with stronger robustness and wider application,a defective pixel repairing method based on deep learning is proposed,which is designed based on the conditional generative adversarial networks(CGAN),and the L1 norm is used as the generation loss function to ensure the accuracy of the generated data.The experimental results show that the ring artifacts are suppressed obviously by the proposed network.