Search

Timber structures have been widely used due to their low-cost and environmental-friendly properties. It is essential to monitor connection damage to ensure the stability and safety of entire timber structures since timber connection damage may induce catastrophic incidents if not detected in a timely manner. However, the current investigations on timber connections focus on mechanical properties and failure modes, and the damage detection of timber connection receives rare attention. Therefore, in this paper, we investigate the damage detection of four common timber connections (i.e., the screw connection, the bolt connection, the decussation connection, and the tooth plate connection) by using the active sensing method. The active sensing method was implemented by using a pair of lead zirconate titanate (PZT) transducers: one PZT patch is used as an actuator to generate stress waves, and the other works as a sensor to detect stress waves after propagating across the timber connection. Based on the wavelet packet energy analysis, the signal energy levels of received stress waves under different damage extent are quantified. Finally, by comparing the signal energy between the intact status and the damage status of the timber connection, we find that the energy attenuates with increasing severity of the connection damage. The experimental results demonstrate that the active sensing method can realize real-time monitoring of timber connection damage, which can guide further investigations. [ABSTRACT FROM AUTHOR]

Image inpainting aims to fill in corrupted regions with visually realistic and semantically plausible contents. In this paper, we propose a progressive image inpainting method, which is based on a forked-then-fused decoder network. A unit called PC-RN, which is the combination of partial convolution and region normalization, serves as the basic component to construct inpainting network. The PC-RN unit can extract useful features from the valid surroundings and can suppress incompleteness-caused interference at the same time. The forked-then-fused decoder network consists of a local reception branch, a long-range attention branch, and a squeeze-and-excitation-based fusing module. Two multi-scale contextual attention modules are deployed into the long-range attention branch for adaptively borrowing features from distant spatial positions. Progressive inpainting strategy allows the attention modules to use the previously filled region to reduce the risk of allocating wrong attention. We conduct extensive experiments on three benchmark databases: Places2, Paris StreetView, and CelebA. Qualitative and quantitative results show that the proposed inpainting model is superior to state-of-the-art works. Moreover, we perform ablation studies to reveal the functionality of each module for the image inpainting task. [ABSTRACT FROM AUTHOR]