A Physics-Informed Neural Network Method for Defect Identification in Polymer Composites Based on Pulsed Thermography

Defect detection in composite materials using active thermography is a well-studied field, and many thermographic data analysis methods have been proposed to facilitate defect visibility enhancement. In this work, we introduce a deep learning method that is constrained by known heat transfer phenomena described by a series of governing equations, also known in the literature as the physics-informed neural network (PINN). The accurate reconstruction of background information based on thermal images facilitates the identification of subsurface defects and reduction in noises caused by an uneven background and heating. The authors illustrate the method’s feasibility through experimental results obtained after pulsed thermography (PT) on a carbon fiber-reinforced polymer (CFRP) specimen.

Lightweight deep network-enabled real-time low-visibility enhancement for promoting vessel detection in maritime video surveillance

Maritime video surveillance has become an essential part of the vessel traffic services system, intended to guarantee vessel traffic safety and security in maritime applications. To make maritime surveillance more feasible and practicable, many intelligent vision-empowered technologies have been developed to automatically detect moving vessels from maritime visual sensing data (i.e., maritime surveillance videos). However, when visual data is collected in a low-visibility environment, the essential optical information is often hidden in the dark, potentially resulting in decreased accuracy of vessel detection. To guarantee reliable vessel detection under low-visibility conditions, the paper proposes a low-visibility enhancement network (termed LVENet) based on Retinex theory to enhance imaging quality in maritime video surveillance. LVENet is a lightweight deep neural network incorporating a depthwise separable convolution. The synthetically-degraded image generation and hybrid loss function are further presented to enhance the robustness and generalisation capacities of LVENet. Both full-reference and no-reference evaluation experiments demonstrate that LVENet could yield comparable or even better visual qualities than other state-of-the-art methods. In addition, it takes LVENet just 0⋅0045 s to restore degraded images with size 1920 × 1080 pixels on an NVIDIA 2080Ti GPU, which can adequately meet real-time requirements. Using LVENet, vessel detection performance can be greatly improved with enhanced visibility under low-light imaging conditions.

Intelligent Vision-Enabled Detection of Water-Surface Targets for Video Surveillance in Maritime Transportation

The timely, automatic, and accurate detection of water-surface targets has received significant attention in intelligent vision-enabled maritime transportation systems. The reliable detection results are also beneficial for water quality monitoring in practical applications. However, the visual image quality is often inevitably degraded due to the poor weather conditions, potentially leading to unsatisfactory target detection results. The degraded images could be restored using state-of-the-art visibility enhancement methods. It is still difficult to generate high-quality detection performance due to the unavoidable loss of details in restored images. To alleviate these limitations, we first investigate the influences of visibility enhancement methods on detection results and then propose a neural network-empowered water-surface target detection framework. A data augmentation strategy, which synthetically simulates the degraded images under different weather conditions, is further presented to promote the generalization and feature representation abilities of our network. The proposed detection performance has the capacity of accurately detecting the water-surface targets under different adverse imaging conditions, e.g., haze, low-lightness, and rain. Experimental results on both synthetic and realistic scenarios have illustrated the effectiveness of the proposed framework in terms of detection accuracy and efficacy.

Examination on the Effectiveness of an Image Processing Technology for the Visibility Enhancement of Firefighters in Fire

In the case of fire, it is important to enhance the visibility of firefighters for emergency activities (for example, fire extinguishment, rescue, and first-aid). In the present study, an image processing technology for visibility enhancement developed by the Korea Electrotechnology Research Institute was used to improve the visibility of firefighters. Image processing technology for visibility enhancement is a technology that combines Multi-Scale Retinex and smoke concentration equalization processing. To examine the effectiveness of the image processing technology in thick smoke conditions, the visibility was classified depending on the smoke generation in an enclosed compartment. In addition, comparative before and after evaluation of image processing technology was performed quantitatively. The visibility was divided into seven levels depending on the recognizable distance of each number plate. Thus, the visibility was improved from a maximum of four levels. Additionally, an in-depth interview was conducted with field crews who are the consumers of this technology; a view of more than 3.5 m was required to use this technology in fire fields.