Acoustic Inspection of Concrete Structures Using Active Weak Supervision and Visual Information

Abstract

Concrete structures are featured heavily in most modern societies. In recent years, the need to inspect those structures has been a growing concern and the automation of inspection methods is highly demanded. Acoustic methods such as the hammering test are one of the most popular non-destructive testing methods for this task. In this paper, an approach to defect detection in concrete structures with active weak supervision and visual information is proposed. Based on audio and position information, pairs of samples are actively queried to a user on their similarity. Those are used to transform the feature space into a favorable one, in a weakly supervised fashion, for clustering defect and non-defect samples, reinforced by position information. Experiments conducted in both laboratory conditions and in field conditions proved the effectiveness of the proposed method.

Wavelet-based multiresolution analysis coupled with deep learning to efficiently monitor cracks in concrete

“…ultrasound) or pull-off methods [24] which are semi-destructive. In the second category, the techniques are much more numerous: we can mention, among others, acoustic methods [25,26] (e.g. acoustic emission, echo, ultrasonic, impact echo), electromagnetic methods [27] (e.g.…”

Weakly Supervised Acoustic Defect Detection in Concrete Structures Using Clustering-Based Augmentation

“…Using the Fast Fourier Transform (FFT), hammering samples are, therefore, first converted from time-series data into Fourier spectrum in a similar fashion as in [4]. Normalization to zero-mean and unit-variance is conducted following the procedure described in [11]. Then, transformation to MFCC is conducted.…”

“…Weakly supervised methods can be thought of as being particularly suited for critical tasks such as inspection work: This is because they allow a subtle mix of human involvement and automation. In [9] and [10] was proposed an initial framework based on weakly supervised metric learning for defect detection in concrete structures, expanded in [11] with the addition of an active query scheme and [12] with the consideration of multiple humans providing the weak supervision. This framework was based on queries to human users on the similarity of selected sample pairs.…”

Ensaios Não Destrutivos Para Diagnóstico De Estruturas De Concreto Armado: Uma Revisão Sistemática

“…Os resultados demonstraram que o esquema proposto obteve uma avaliação de precisão favorável com alta eficiência e baixa carga de computação. Segundo Louhi Kasahara et al (2020), uma abordagem ativa fracamente supervisionada e reforçada por informação visual para a inspeção de estruturas de concreto usando métodos acústicos foi utilizada para detecção de defeitos em estruturas de concreto. Experiências em condições de laboratório, utilizando blocos de teste de concreto com defeitos provocados de delaminações e vazios, e em condições de campo, utilizando túnel simulado e módulo de martelamento automatizado foi realizado.…”