Combination and automation of computer-aided methods for quantitative crack analysis of concrete - Part 2

薄切片的2D图像分析逐渐自动化并得到验证。除此之外，还可以在预处理和分析过程中实现优化。一个类似的方法已被用于使用计算机断层扫描的标本的3D分析。为了获得更多的数据，这些2D和3D图像分析相互结合。在量化整个样品体积方面还不能取得令人满意的结果，但其可行性得到了证明。该项目的进一步工作将使这最后一步成为可能。继续进行的研究项目分为三个阶段。第三阶段将继续进行并完成。该阶段包括最终量化试样的总裂纹体积以及2D和3D图像分析的比较。从后者的关键信息是通过结合这两种方法。此外，还制作了测试样品，以测试组合方法在实践中的应用。因此，这种方法的局限性和优点可以以一种可理解的方式来展示。在第四阶段，将根据裂缝在混凝土基体中的位置对裂缝进行自动分类。这将使得能够获得关于裂缝的各种原因的更有效的结果，例如冻结和融化、碱-硅反应或收缩。在第五阶段，在开展上述工作的同时，将对迄今为止制定的工作流程进行反复调整。此外，目的是为科学界公布工作流程。该方法将免费提供，包括说明、应用示例和源代码。

The 2D image analysis of thin sections was gradually automated and validated. Apart from this, optimisations could be realised regarding pre-processing and in the analysis process. A comparable approach has been adopted for the 3D analysis of specimens using computer tomography. To acquire further data, these 2D and 3D image analyses were combined with each other. Satisfactory results in quantifying the entire specimen volume could not yet be achieved, but their feasibility was demonstrated. Further work on the project will make this final step possible. The research project to be continued is divided into three phases. Phase III will be continued and completed. The phase includes the final quantification of the total crack volume of a specimen as well as the comparison of 2D and 3D image analysis. From the latter the key information is obtained by combining the two methods. Further on, test specimens are produced which test the application of the combined methods in practice. As a result, the limits and advantages of this method can be demonstrated in a comprehensible way. In phase IV, an automated classification of the cracks based on their location in the concrete matrix will be implemented. This will make it possible to derive more efficient results with regard to the various causes of cracks, such as freezing and thawing, the alkali-silica reaction or shrinkage. In phase V, the work described above will be accompanied by an iterative adaptation of the workflow developed so far. In addition, the aim is to publish the workflow for the scientific community. The method will be made freely accessible, including a description, application examples and source code.