CN-121999095-A - CT image ring artifact correction method based on detector jitter

CN121999095ACN 121999095 ACN121999095 ACN 121999095ACN-121999095-A

Abstract

The invention relates to the technical field of CT artifact correction and image processing, in particular to a CT image ring artifact correction method based on detector shake, which controls a detector to shake according to a preset periodic shake track and collect a projection image during scanning; the method comprises the steps of correcting an image to be in a non-jittering state according to the track, carrying out CT reconstruction on the corrected image, and carrying out bilateral filtering on the reconstructed CT image to obtain a corrected image. The jitter track follows a triangular wave function, bilinear interpolation is adopted for return, FDK algorithm is adopted for reconstruction, and Gaussian weighting mode combining space and pixel value domain is adopted for bilateral filtering. The invention breaks the ring artifact generation mechanism through physical dithering, combines edge protective filtering, effectively removes ring artifacts, maintains image definition, and has the advantages of no dependence on extra correction, no need of a large amount of training data and simple implementation.

Inventors

LIAO JUAN
DAI CHAOWEI
LIU PAN
ZHENG PEICHEN
CHENG SHUGANG
YE JUNCHAO

Assignees

重庆日联科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (8)

1. The CT image ring artifact correction method based on detector dithering is characterized by comprising the following steps of: in the process of scanning a workpiece, the detector is controlled to shake according to a preset periodic shaking track so as to acquire a workpiece projection image; correcting the projection image to an unjittered state according to the recorded jittering track; performing CT reconstruction on the corrected projection image to obtain a CT image; And carrying out bilateral filtering treatment on the CT image to obtain an image after ring artifact correction.
2. The detector dithering-based CT image ring artifact correction method as recited in claim 1, wherein, in dithering the control detector according to a preset periodic dithering trajectory: the periodic dither trajectory follows a triangular wave function.
3. The detector jitter based CT image ring artifact correction method as recited in claim 2, wherein said triangular wave function is expressed as: Wherein, the For the amplitude of the jitter to be high, For the period of the jitter period, Is the serial number of the projected image.
4. The detector-dithering-based CT image ring artifact correction method as recited in claim 1, wherein, in said returning the projection image to the un-dithered state: and correcting the projection image to the non-jittered state by adopting a bilinear interpolation method.
5. The detector-jitter-based CT image ring artifact correction method as recited in claim 4, wherein said bilinear interpolation method is expressed as: Wherein, the The coordinates of the four pixel points are And The corresponding pixel value is And 。
6. The detector-jitter-based CT image ring artifact correction method as recited in claim 1, wherein in said CT reconstruction of said corrected projection image: the CT reconstruction adopts an FDK algorithm.
7. The detector-jitter-based CT image ring artifact correction method as recited in claim 1, wherein, in said bilateral filtering of said CT image: the bilateral filtering processing adopts a Gaussian weighting mode of combining a spatial domain and a pixel value domain, and the expression is as follows: Wherein, the In order to input a CT image, In order to filter the image after it has been processed, In order to filter the radius of the window, As a result of the normalization factor, And The spatial weight function and the pixel weight function are respectively, and are Gaussian functions.
8. The detector dithering-based CT image ringing correction method of claim 7, wherein the spatial weighting function And pixel weight function Standard deviations of (2) are respectively And 。

Description

CT image ring artifact correction method based on detector jitter Technical Field The invention relates to the technical field of CT artifact correction and image processing, in particular to a CT image ring artifact correction method based on detector shake. Background X-ray computed tomography (Computed Tomography, CT) is taken as a nondestructive testing tool, can clearly show the three-dimensional structure inside a workpiece, and has become a pillar technology for quality detection and analysis in industries such as aerospace, automobile manufacturing, electronic semiconductors and the like. The quality of the CT image determines the accuracy of the detection. In actual detection, the inconsistent response of the detector pixels can lead to ring artifacts on the CT image. When the ring artifact overlaps with the defect, the recognition is seriously disturbed, resulting in false detection or missing detection. Therefore, the ring artifact correction is important to improving the CT image quality and guaranteeing the detection reliability. Currently, three main methods for ring artifact correction are detector correction, analysis iteration and deep learning. The detector correction method needs to interrupt the normal scanning flow, and the response curves of the pixels of the detector tend to be consistent by executing dark field, bad pixels, bright field correction and other means, and then scanning is continued. However, the corrected response curve may fail again in subsequent scans due to factors such as instability of the source output, random noise of the detector, and the like. The analytic iteration method adopts operations such as filtering, interpolation, transformation, iteration and the like to remove ring artifacts in CT images, but image details are easy to lose, so that the images are blurred. Deep learning is data driven, requires a large amount of data to learn to remove ring artifacts, the correction effect depends on the quality of the data set, and the training time is long. Disclosure of Invention The invention aims to provide a CT image ring artifact correction method based on detector shake, which solves the problems that the existing detector correction method is likely to fail, an analysis iteration method is easy to cause image detail loss and blurring, and a deep learning method depends on a large amount of data and has long training time. In order to achieve the above object, the present invention provides a method for correcting CT image ring artifact based on detector shake, comprising the steps of: in the process of scanning a workpiece, the detector is controlled to shake according to a preset periodic shaking track so as to acquire a workpiece projection image; correcting the projection image to an unjittered state according to the recorded jittering track; performing CT reconstruction on the corrected projection image to obtain a CT image; And carrying out bilateral filtering treatment on the CT image to obtain an image after ring artifact correction. Wherein, in the process that the control detector shakes according to a preset periodic shaking track: the periodic dither trajectory follows a triangular wave function. Wherein the triangular wave function is expressed as: Wherein, the For the amplitude of the jitter to be high,For the period of the jitter period,Is the serial number of the projected image. Wherein, in said returning the projected image to the non-dithered state: and correcting the projection image to the non-jittered state by adopting a bilinear interpolation method. Wherein, the bilinear interpolation method is expressed as: Wherein, the The coordinates of the four pixel points areAndThe corresponding pixel value isAnd。 Wherein, in the CT reconstruction of the projection image after the alignment: the CT reconstruction adopts an FDK algorithm. Wherein, in the bilateral filtering processing of the CT image: the bilateral filtering processing adopts a Gaussian weighting mode of combining a spatial domain and a pixel value domain, and the expression is as follows: Wherein, the In order to input a CT image,In order to filter the image after it has been processed,In order to filter the radius of the window,As a result of the normalization factor,AndThe spatial weight function and the pixel weight function are respectively, and are Gaussian functions. Wherein the spatial weight functionAnd pixel weight functionStandard deviations of (2) are respectivelyAnd。 The invention discloses a detector jitter-based CT image ring artifact correction method, which belongs to the field of CT artifact correction and image processing, and aims to remove ring artifacts by breaking a ring forming mechanism by using a jitter detector and combining a bilateral filter from the ring artifact principle. Compared with three types of ring artifact correction methods in the prior art, the method has the advantages that: 1. the image quality is high, and the image blurring can not