Image Alignment and Flat-Field Correction of Film and Computed Radiography Images on the LLNL Flash Testbed

Wilson, Finnegan, Harke, Kathryn, Hardy, Anthony et al. · 2025

Read Image Alignment and Flat-Field Correction of Film and Computed Radiography Images on the LLNL Flash Testbed by Wilson, Finnegan, Harke, Kathryn, Hardy, Anthony et al. (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/doe-osti-2587720-image-alignment-and-flat-field-correctio.txt.

Introduction

Finnegan Wilson leads a six-author team in this September 2025 report, LLNL-TR-2010240, from Lawrence Livermore National Laboratory. Kathryn Harke, Anthony Hardy, Kyle Sullivan, Veronica Eliasson, and Joseph Tringe complete the team. Their subject is cleaning up flash X-ray images taken on the laboratory's Flash Testbed, where film and computed radiography plates record shots inside firing tanks.

A standard fix is flat-field correction: divide the experiment image by a matching blank image, canceling detector unevenness. The catch is alignment, since the plate shifts between the two shots. The team's workflow uses fiducial marks from small stainless steel ball bearings, 3.18 millimeters across, planted in the scene. Software windows walk the user through picking the fiducial centers, computing the shift, and applying the correction to the second image. The report documents the procedure for the groups that shoot radiographs at the site.

Catalog record

CREATORWilson, Finnegan, Harke, Kathryn, Hardy, Anthony et al.
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE19 KB
LICENSEPublic Domain
RIGHTSPUBLIC DOMAIN
SOURCEDOE OSTI

Source: https://www.osti.gov/biblio/2587720

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

LLNL-TR-2010240 Image Alignment and Flat-Field Correction of Film and Computed Radiography Images on the LLNL Flash Testbed F Wilson, K Harke, A Hardy, K Sullivan, V Eliasson, J Tringe September 2025 This document was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or Lawrence Livermore National Security, LLC, and shall not be used for advertising or product endorsement purposes. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-TR-2010240 Image Alignment and Flat-Field Correction of Film and Computed Radiography Images on the LLNL Flash Testbed F Wilson, K Harke, A Hardy, K Sullivan, V Eliasson, J Tringe September 2025 This document was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or Lawrence Livermore National Security, LLC, and shall not be used for advertising or product endorsement purposes. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. Image Alignment and Flat-Field Correction of Film and Computed Radiography Images on the LLNL Flash Testbed Finnegan Wilson Kathryn Harke Anthony Hardy Kyle Sullivan Veronica Eliasson Joseph Tringe 1. Introduction Computed radiography (CR) imaging plates and film are used in HEAF firing tanks and in the NDE group. The imaging plates allow for the creation of high-resolution digital images with flash X-ray (FXR). A typical treatment of flash X-ray radiographs is flat-field correction, where the image from an experiment is normalized by a “flat-field” or “bright- field” image. This flat-field image is taken in an identical configuration to the experimental image, but without the object or experiment in the field of view (FOV). This treatment reduces spatial effects from the FXR spot size and detector misalignment, as each pixel value in the corrected image represents a ratio of collected radiation with the object in FOV to the collected radiation without the object in FOV. One challenge in the creation of a corrected image is the misalignment of the CR plate or film pack between capturing the flat-field image and the experimental image. Fiducial structures can remedy this issue. Small (3.18mm) stainless steel ball bearings serve as good fiducial structures due to their small size and high radiographic contrast. Spheres are view-agnostic geometry, always presenting a circular cross section no matter the orientation. This process was developed for images from the flash X-ray testbed. The flash testbed was used to compare the X-ray transmission at different thicknesses of aluminum and copper step wedges. Each step wedge section is a rectangular shape, so evaluation is made much simpler if the rectangles are not rotated with respect to the image. This alignment process aligns the object and flat-field images together and leaves the rectangles of each step aligned with the image. 1 2. Fiducial-Based Image Alignment 2.1 Software This image correction script runs in MATLAB and was written in version R2022b. The program relies on functions from the Image Processing Toolbox, downloadable from MathWorks or directly within the MATLAB application. The script will be included in the appendix of this report. Indexing and variable names are in reference to a broader workflow and should be adjusted to fit the user’s application. 2.2 Process overview This image alignment process relies on two circular fiducials included in the FOV in both the experimental image and in the flat-field image. Figure 1 (a) shows an image of a stepped wedge and line pair gauge taken on film and scanned directly into a digital format. At the bottom face of the steps wedge, two fiducial spheres can be seen on the left and right of the wedge. Figure 1 (b) shows the flat-field image to highlight that the fiducial structures are still in the FOV. (a) (b) Figure 1: Example film image. (a) shows the “experimental image” with objects and fiducial spheres in FOV, and (b) shows the flat-field image without objects. 2 In between taking image (a) and image (b), the step wedge and line pair gauge were removed from the alignment stand. The stand maintains alignment between the objects and the fiducials. The stand is also fixed in place relative to the center of the X-ray tube. The film pack is removed, processed, and replaced with a fresh film between shots. The film pack is replaced at the same location, but small rotations of the film pack were not controlled for. 2.3 Getting started The first step to running this script is to upload the raw images to MATLAB. The “imread” function works well to import a TIF file directly. Variable labels “im0” and “bk0” refer to the experimental image file and the background/flat-field image file, respectively. The script shown in the appendix references a MATLAB structure called “B0” with the set of TIF images saved directly in MATLAB, and index “i” refers to image number within the set. For the data set used in this example, the background shot was captured directly after each experimental image, resulting in indexing of “i” for “im0” and “i+1” for “bk0”. There are two additional inputs to the alignment script: The first is “bit-depth”, where the digital bit depth of the scanned image can be input. This value is only used for scanned film images, which are uploaded as “attenuation” images, where bright areas represent more attenuation of the incident radiation. Scanned CR images are in an inverted format, with bright regions representing more transmission of the incident beam. The second input is the “gain” required to view the raw images. Depending on the imaging set up, the raw scanned image may appear too dark to see without scaling. By adding a gain value, for example 10, the image is artificially brightened when displayed for user input. The second section of the script, used to create a cropped and flat-field

Read the complete text