Detection of the liquefied natural gas (LNG) leakage attracts increasing attention for preventing environments and governments from severe pollution and economic loss. Existing frameworks take advantage of stationary surveillance thermal cameras to detect the LNG leakage, which comprises background subtraction and leakage classification. However, these methods are limited in rural areas due to the lack of sensitivity and accuracy. In this article, a generalized framework, i.e., tensor-based leakage detection (TBLD), is proposed to detect LNG leakage in the rural area from surveillance thermal cameras. First, the proposed TBLD takes advantage of tensor factorization to fuse thermal image and corresponding gradient maps for improving sensitivity. Additionally, a finite-state-machine is designed to maintain leakage foreground along with the video streaming. The experiments demonstrate the robust performance of TBLD in the background subtraction stage. Second, multiple classification techniques are explored in the leakage classification stage. The results suggest that the TBLD can accurately detect the LNG leakage by applying 50 layers of residual networks (ResNet50). Finally, compared with contemporary frameworks, the TBLD has consistently improved performance concerning the different distances of LNG leakage. The experimental results demonstrate the effectiveness of the proposed TBLD, which also shows the great potential of TBLD in future industrial applications.

Surface defects directly affect the mechanical properties of industrial strip steel products. To evaluate the integrity of the strip steel surface, a channel-wise global Transformer based dual-branch network for strip steel surface defect detection, dubbed CGTD-Net, is proposed in this study. First, the strip steel surface images are pre-processed using saturation adjustment and random flipping strategies to remove unnecessary background information and improve network generalization. Second, the Swin Transformer is employed at the end of the backbone network and the negative impacts of a single channel are then mitigated by using the multi-channel feature pyramid networks via Transformer, which improves the extraction ability of the global semantic information for tiny or narrow defects. Third, an edge detection branch network is constructed with a spatial-channel attention module (SCGA) to further enhance the feature extraction on both spatial and channel information. Finally, the CGTD-Net is compared with eleven state-of-the-art methods on the NEU-DET dataset, and ablation experiments are also implemented. The comparison results, conducted on a single 3090Ti GPU, reveal that the CGTD-Net achieves a mean intersection over union (mIoU) of 75.16% at 178 frames per second (FPS), outperforming other methods. The ablation experiment demonstrates that the CGTD-Net improves the mIoU by 7.83% and the F-score by 6.3% compared to the baseline.

The intelligent fault diagnosis powered deep learning (DL) is widely applied in various practical industries, but the conventional intelligent fault diagnosis methods cannot fully juggle the manifold structure information with multiple-order similarity from the massive unlabeled industrial data. Thus, a new Multiple-Order Graphical Deep Extreme Learning Machine (MGDELM) algorithm for unsupervised fault diagnosis of rolling bearing is proposed in this study. Specifically, the developed MGDELM algorithm mainly contains two parts: i). one is unsupervised multiple-order feature extraction, the first-order proximity with Cauchy graph embedded is applied to extract the local structural information, and the second-order proximity is simultaneously employed for mining global structural information; ii). the other is used the unsupervised Fuzzy-C-Mean (FCM) into fault clustering built on the extracted multiple-order graph embedding features. Empirically, two cases of rolling bearing failure data validate the effectiveness of the proposed algorithm and fault diagnosis method. By jointly optimizing the multiple-order objective function, the proposed MGDELM algorithm can synchronously extract local and global structural information from the raw industrial data. This study also provides a novel promising approach for unsupervised fault diagnosis.

A leak detection and repair survey (LDAR) is essential to ensure a reliable and safe liquefied natural gas (LNG) supply. Modern LDAR systems deploy numerous fixed thermal imaging devices to automatically monitor the risk of potential leaks empowered by computational intelligence frameworks. Existing frameworks employ either background subtraction-based (BGS-based) or deep neural network-based (DNN-based) frameworks for LNG leak detection from thermal images. However, the BGS-based frameworks feature high sensitivity to perceive LNG emissions with low precision. On the contrary, the DNN-based frameworks can precisely classify the LNG leak after training while the sensitivity is low. Additionally, conventional DNN-based frameworks are difficult in modeling non-rigid objects such as LNG gas due to limited perceptive fields. Therefore, this study proposes a hybrid framework, namely foreground fusion-based gas detection (FFBGD), combining the advantages of BGS-based and DNN-based detectors for improved detection robustness through newly introduced concept of information fusion to LNG industries. Specifically, a foreground fusion network (FFN) is designed to fuse information of original thermal and foreground images after BGS based on the visual attention mechanism. Meanwhile, several advanced modules, i.e. deformable convolution, feature pyramid network, and cascade region-of-interest (ROI) head are adopted to enhance leak detection by offering better perceptive fields. Extensive experiments are carried out in this study to demonstrate the significant improvement brought by the proposed FFBGD over leak detection accuracy and robustness. Hence, the proposed solution can be deployed in energy facilities and enable reliable visual surveillance of LNG leaks.

The proton precession magnetometer (PPM) is a commonly used device to measure the varying magnetic field. Since the frequency of the PPM sensing free induction decay (FID) signal is proportional to the magnetic field, the signal-to-noise ratio (SNR) is always a critical issue that influences the measurement accuracy severely due to the external interferences such as harmonic noise and random noise. In this study, to boost the SNR of the FID signal, an effective filtering algorithm based on time-frequency peak filtering (TFPF) analyzed with pseudo-Wigner–Ville distribution (PWVD) is proposed. Through pre-treating the collected noisy FID signal with frequency modulation and instantaneous frequency estimation using the peak value of the time-frequency characterization, the embedded noise can be decorrelated and the relative pure FID signal can be detected regardless of the impact of varying noise levels. The superiority of the proposed synaptic noise reduction framework, namely, TFPF-PWVD, was found by comparing it with state-of-the-art approaches under the same conditions. The results illustrated that even though in a strong-noisy scenario, the proposed TFPF-PWVD based approach still achieved the best SNR for the yielded sensing FID and the minimum standard deviation for the observed magnetic field data, which can enhance the geomagnetic measuring performance of a PPM.