Electroluminescence (EL) imaging is a widely adopted method in quality assurance of the photovoltaic (PV) manufacturing industry. With the growing demand for high-quality PV products, automatic inspection methods based on machine vision have become an emerging area concern to replace manual inspectors. Therefore, this paper presents an automatic defect-inspection method for multi-cell monocrystalline PV modules with EL images. A processing routine is designed to extract the defect features of the PV module, eliminating the influence of the intrinsic structural features. Spectrum domain analysis is applied to effectively reconstruct an improved PV layout from a defective one by spectrum filtering in a certain direction. The reconstructed image is used to segment the PV module into cells and slices. Based on the segmentation, defect detection is carried out on individual cells or slices to detect cracks, breaks, and speckles. Robust performance has been achieved from experiments on many samples with varying illumination conditions and defect shapes/sizes, which shows the proposed method can efficiently distinguish intrinsic structural features from the defect features, enabling precise and speedy defect detections on multi-cell PV modules.

Robotics is widely used in nearly all sorts of manufacturing. Steady performance and accurate movement of robotics are vital in quality control. Along with the coming of the Industry 4.0 era, oceans of sensor data from robotics are available, within which the health condition and faults are enclosed. Considering the growing complexity of the manufacturing system, an automatic and intelligent health-monitoring system is required to detect abnormalities of robotics in real-time to promote quality and reduce safety risks. Therefore, in this study, we designed a novel semantic-based modeling method for multistage robotic systems. Experiments show that sole modeling is not sufficient for multiple stages. We propose a descriptor to conclude the stages of robotic systems by learning from operational data. The descriptors are akin to a vocabulary of the systems; hence, semantic checking can be carried out to monitor the correctness of operations. Furthermore, the stage classification and its semantics were used to apply various regression models to each stage to monitor the quality of each operation. The proposed method was applied to a photovoltaic manufacturing system. Benchmarks on production datasets from actual factories show the effectiveness of the proposed method to realize an AI-enabled real-time health-monitoring system of robotics.

In camera distortion correction and calibration research, there is a prevalent assumption that distortion exhibits radial symmetry around a center of distortion. However, the validity of this assumption is contingent upon the extent of tangential distortion in the system. This study delves into a quantitative analysis of tangential distortion by simulating diverse camera specifications and tilt angles between lenses and sensor arrays. The findings indicate that larger field-of-view and smaller pixel sizes lead to an augmentation of tangential distortion. In practical camera settings, the tangential distortion can extend beyond a dozen pixels. Consequently, traditional models relying on radial symmetry may encounter a decline in precision when addressing such distortions. Moreover, the study suggests that centers of distortion determined geometrically based on radial symmetry may not be numerically optimal in scenarios with tangential distortions. As a result, the robustness of the radial symmetry assumption may be compromised in the face of prominent trends in camera advancements.

The ease of using sensors and collecting data provides a great support for industry reliability. Rolling bearings are indispensable components in various machinery systems, and accurately predicting their remaining useful life (RUL) is crucial for ensuring equipment safety and reliability through prognostic health maintenance (PHM). While traditional methods primarily rely on analyzing vibration signals for degradation modeling, they often miss out on crucial insights embedded in planar vibration patterns. In our research, we propose a groundbreaking approach that utilizes planar vibration analysis to capture a comprehensive range of degradation information. We introduce planar vibration features (PVF) which are derived from analyzing the scattering of acceleration vectors on a 2D plane. These PVFs reveal significant correlations with RUL. Meanwhile, unlike conventional methods that rely on single timestamps or windowed samples, our proposed model harnesses the power of multilayer convolutional neural networks (CNNs) to learn from the entire degradation history. To tackle the challenge of early-stage uncertainty, we have devised a weighted mean square error metric. Through validation on the PHM2012 challenge dataset, our approach demonstrates superior accuracy compared to recent methods for RUL prediction. This research sheds light on the importance of capturing multidirectional vibration information for effective RUL prediction, thereby offering a promising avenue for advancing machinery prognostics and maintenance practices.

This paper presents a robust, decoupled approach to camera distortion correction using a rational function model (RFM), designed to address challenges in accuracy and flexibility within precision-critical applications. Camera distortion is a pervasive issue in fields such as medical imaging, robotics, and 3D reconstruction, where high fidelity and geometric accuracy are crucial. Traditional distortion correction methods rely on radial-symmetry-based models, which have limited precision under tangential distortion and require nonlinear optimization. In contrast, general models do not rely on radial symmetry geometry and are theoretically generalizable to various sources of distortion. There exists a gap between the theoretical precision advantage of the Rational Function Model (RFM) and its practical applicability in real-world scenarios. This gap arises from uncertainties regarding the model’s robustness to noise, the impact of sparse sample distributions, and its generalizability out of the training sample range. In this paper, we provide a mathematical interpretation of how RFM is suitable for the distortion correction problem through sensitivity analysis. The precision and robustness of RFM are evaluated through synthetic and real-world experiments, considering distortion level, noise level, and sample distribution. Moreover, a practical and accurate decoupled distortion correction method is proposed using just a single captured image of a chessboard pattern. The correction performance is compared with the current state-of-the-art using camera calibration, and experimental results indicate that more precise distortion correction can enhance the overall accuracy of camera calibration. In summary, this decoupled RFM-based distortion correction approach provides a flexible, high-precision solution for applications requiring minimal calibration steps and reliable geometric accuracy, establishing a foundation for distortion-free imaging and simplified camera models in precision-driven computer vision tasks.