The core problem in process tomography is that the image reconstruction process is ill-posed which limits the image quality. Dual-modality tomography is effective to acquire more information in the sensing area, which mitigates the ill-posedness of image reconstruction. However, dual-modality measurement data are of different physical dimensions and thus difficult to fuse together which presents a great challenge. To effectively use the dual-modality information, an ERT and UTT projection sorting (EUPS) dual-modality image reconstruction algorithm is presented. The EUPS algorithm converted the dual-modality information to a normalized projection parameter. The physical concept of the proposed parameter is the contribution that each imaging pixel projected on the measurement data. The proposed normalized parameter is dimensionless which can be fused together between different modalities. The EUPS algorithm was experimentally validated by several models comparing with several classical algorithms. The results indicated that the image reconstruction quality of the proposed dual-modality algorithm is superior to single-modality image reconstruction algorithms.

中文翻译：

---

ERT和UTT投影排序算法的双模态断层扫描

过程断层扫描的核心问题是图像重建过程是不适定的，这限制了图像质量。双模态断层扫描可以有效地在传感区域获取更多信息，从而减轻图像重建的不适定性。然而，双模态测量数据具有不同的物理尺寸，因此难以融合在一起，这是一个很大的挑战。为了有效利用双模态信息，提出了一种ERT和UTT投影排序（EUPS）双模态图像重建算法。EUS 算法将双模态信息转换为归一化的投影参数。提出的参数的物理概念是每个成像像素投影到测量数据上的贡献。建议的归一化参数是无量纲的，可以在不同的模态之间融合在一起。与几种经典算法相比，EUPS 算法通过几种模型进行了实验验证。结果表明，所提出的双模态图像重建算法的图像重建质量优于单模态图像重建算法。