Automated Measurement of Geometric Features in Curvilinear Structures Exploiting Steger’s Algorithm

Abstract

Accurately assessing the geometric features of curvilinear structures on images is of paramount importance in many vision-based measurement systems targeting technological fields such as quality control, defect analysis, biomedical, aerial, and satellite imaging. This paper aims at laying the basis for the development of fully automated vision-based measurement systems targeting the measurement of elements that can be treated as curvilinear structures in the resulting image, such as cracks in concrete elements. In particular, the goal is to overcome the limitation of exploiting the well-known Steger’s ridge detection algorithm in these applications because of the manual identification of the input parameters characterizing the algorithm, which are preventing its extensive use in the measurement field. This paper proposes an approach to make the selection phase of these input parameters fully automated. The metrological performance of the proposed approach is discussed. The method is demonstrated on both synthesized and experimental data.

In the Direction of an Artificial Intelligence-Enabled Monitoring Platform for Concrete Structures

“…The measurement of the crack aperture was performed using an AI model trained on a real concrete dataset, able to segment the material failure, hence localizing the defect in combination with an algorithm based on Steger’s theory [ 36 ] for the detection and measurement of curvilinear structures. To better identify the defects, it is necessary to train the neural network with a consistent dataset of real-life images; hence, a dataset was created with real images of cracks in concrete gathered both from websites and in-field.…”

“…To better identify the defects, it is necessary to train the neural network with a consistent dataset of real-life images; hence, a dataset was created with real images of cracks in concrete gathered both from websites and in-field. The neural network used for segmentation and defect identification is UNet [ 36 ]. It is worthy to underline that the proposed approach allows the identification of a crack on the picture of the cement-based element, to precisely locate it with sub-pixel resolution, and to assess its aperture width [ 36 ].…”

“…The neural network used for segmentation and defect identification is UNet [ 36 ]. It is worthy to underline that the proposed approach allows the identification of a crack on the picture of the cement-based element, to precisely locate it with sub-pixel resolution, and to assess its aperture width [ 36 ]. The visible images were employed for the detection and quantification of the crack aperture, utilizing the image reference system expressed in pixels.…”