This paper proposes an automated approach to predict crack pattern similarities that correlate well with assessment by structural engineers. We use Siamese convolutional neural networks (SCNN) that take two crack pattern images as inputs and output scalar similarity measures. We focus on 2D masonry facades with and without openings. The image pairs are generated using a statistics-based approach and labelled by 28 structural engineering experts. When the data is randomly split into fit and test data, the SCNNs can achieve good performance on the test data ((Formula presented.)). When the SCNNs are tested on ”unseen” archetypes, their test (Formula presented.) values are on average 1% lower than the case where all archetypes are ”seen” during the training. These very good results indicate that SCNNs can generalise to unseen cases without compromising their performance. Although the analyses are restricted to the considered synthetic images, the results are promising and the approach is general.

@en

The degree of similarity between damage patterns often correlates with the likelihood of having similar damage causes. Therefore, deciding whether crack patterns are similar is one of the key steps in assessing the conditions of masonry structures. To our knowledge, no literature has been published regarding masonry crack pattern similarity measures that would correlate well with assessment by structural engineers. Hence, currently, similarity assessments are solely performed by experts and require considerable time and effort. Moreover, it is expensive, limited by the availability of experts, and yields only qualitative answers. In this work, we propose an automated approach that has the potential to overcome the above shortcomings and perform comparably with experts. At its core is a deep neural network embedding that can be used to calculate a numerical distance between crack patterns on comparable façades. The embedding is obtained from fitting a deep neural network to perform a classification task; i.e., to predict the crack pattern archetype label from a crack pattern image. The network is fitted to synthetic crack patterns simulated using a statistics-based approach proposed in this work. The simulation process can account for important crack pattern characteristics such as crack location, orientation, and length. The embedding transforms a crack pattern (raster image) into a 64-dimensional real-valued vector space where the closeness between two vectors is calculated as the cosine of their angle. The proposed approach is tested on 2D façades with and without openings, and with synthetic crack patterns that consist of a single crack and multiple cracks.@en

There is a paradigm shift from two- to three-dimensional data, from maps to information dense models. Self-driving cars, digitization of historic buildings or maintenance of highway infrastructure are a small selection of many applications that use laser scanning to acquire three-dimensional data of our physical surroundings. Most of these applications require more than the shape of their surroundings. For example, a self-driving car needs to identify pedestrians, road signs and traffic lights in order to navigate safely. Therefore the point cloud acquired by laser scanning needs to be enriched with additional information. Automatic assignment of the object type a point belongs to is called classification.

This research focuses on deep learning for point cloud classification, because it revolutionized classification of imagery. PointNet is used to enable deep learning directly on point cloud data sets. To date PointNet is proven for indoor point clouds, this research explores the application of PointNet on an outdoor highway scene. The methodology creates point cloud training data efficiently by reusing known 2D object locations. Different spatial representations and sampling methods for the points are tested.

On average the method classifies 50% of the points correctly in four object classes. In combination with clustering of the point-wise predictions, the method predicts 60% of the 2D object locations successfully. The performance is comparable to the 47% average class accuracy PointNet achieves for 13 classes on the indoor data set.