A fast iterative reconstruction method of sparse angle CT for cylindrical lithium battery

Sparse-angle CT (Computer Tomography) imaging can effectively reduce the radiation exposure risk to lithium batteries. This study designs an iterative reconstruction algorithm for the reconstruction problem of cylindrical lithium batteries using SART (Simultaneous Algebraic Reconstruction Technique) as the fidelity term and tangential smoothing filtering as the regularization term. Additionally, after performing diagonal interpolation on the original sinogram followed by an inverse Radon transform, the obtained slice image is used as initial values to accelerate the iterative process; a weighted template is designed based on the sparsity of different image regions to suppress the smoothing effect of the regularization term on image details, thus improving the signal-to-noise ratio of the reconstructed images. Experiments on simulated images, real X-ray images, and neutron images, including sparsity experiments, noise resistance experiments, ablation experiments, and comparative experiments, are conducted using peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and mean square error (MSE) as image quality metrics, validating the effectiveness, noise resistance, and superiority of the proposed method under high sparsity conditions. For cylindrical lithium batteries, the proposed method can effectively suppress sparse artifacts in slice images, better restoring the structure and details of the images. The algorithm has low complexity, fast convergence speed, and high engineering application value.