As an important driving device, the permanent magnet synchronous motor (PMSM) plays a critical role in modern industrial fields. Given the harsh working environment, research into accurate PMSM fault diagnosis methods is of practical significance. Time–frequency analysis captures the rich features of PMSM operating conditions, and convolutional neural networks (CNNs) offer excellent feature extraction capabilities. This study proposes an intelligent fault diagnosis method based on continuous wavelet transform (CWT) and CNNs. Initially, a mechanism analysis is conducted on the inter-turn short-circuit and demagnetization faults of PMSMs, identifying and displaying the key feature frequency range in a time–frequency format. Subsequently, a CNN model is developed to extract and classify these time–frequency images. The feature extraction and diagnosis results are visualized with t-distributed stochastic neighbor embedding (t-SNE). The results demonstrate that our method achieves an accuracy rate of over 98.6% for inter-turn short-circuit and demagnetization faults in PMSMs of various severities.

Vibration measurement serves as the basis for various engineering practices such as natural frequency or resonant frequency estimation. As image acquisition devices become cheaper and faster, vibration measurement and frequency estimation through image sequence analysis continue to receive increasing attention. In the conventional photogrammetry and optical methods of frequency measurement, vibration signals are first extracted before implementing the vibration frequency analysis algorithm. In this work, we demonstrate that frequency prediction can be achieved using a single feed-forward convolutional neural network. The proposed method is verified using a vibration signal generator and excitation system, and the result compared with that of an industrial contact vibrometer in a real application. Our experimental results demonstrate that the proposed method can achieve acceptable prediction accuracy even in unfavorable field conditions.

An $n$ -order marker field is a special binary matrix whose $n\times n$ subregions are all distinct from each other in four orientations. It is commonly used to guide the composing process of position-sensing markers, which can be detected and identified in a camera image with very limited scope or severe visibility problems. Despite the advantages, position-sensing markers are rare and overlooked because generating marker fields is difficult. In this paper, we broaden the definition of marker field, making it more powerful and flexible. Then, we propose bWFC (binary wave function collapse) and its high-speed version, fast-bWFC, to solve the generation problem. The methods are packaged into an open-sourced toolkit named HydraMarker, with which, users not only can generate marker fields on laptops within a short period of time, but also can highly customize them: preset values; fields and subregions in any shape; multifold local uniqueness. Comparative results indicate that the proposed method has superior efficiency, quality, and capability. It makes marker field generation accessible to common marker designers, opening up more possibilities for fiducial markers.

As image acquisition devices have outstanding potential for gathering vibration information, computer vision has received a lot of interest in structural health monitoring (SHM). In this work, a fully automated peak picking methodology based on computer vision in tandem with deep learning is proposed to realize vibration measurements and identify natural frequencies from the plot of the power spectral density transmissibility (PSDT). A deep-learning-enhanced image processing technology was used to extract the vibration signals with automatic active pixel selection, while a convolutional neural network was used to further process the vibration measurements so that the frequencies could be identified from PSDT-based functions. The proposed method was verified by three case studies, including the dynamic testing of two laboratory models and the field testing of the stay cable. The findings showed that the proposed deep-learning-enhanced approach has a high potential for use in SHM by automatically performing vibration measurement and frequency extraction.

Dark channel prior is a highly valued theory mostly used in the area of haze removal. It reveals the wide existence of pixels with very small channel. However, the definition is in a decorated way which relies on the concept of dark channel. The dark channel works like an image domain wherein non-sky regions of clear outdoor images are close to zero. In this paper, we show that such a concept is unnecessary, misleading, and can be improved by using a more basic concept named dark pixel. A novel transmission estimation method based on finding dark pixels is proposed. The method is concise and easy to understand. In comparisons with state-of-the-art methods, our method shows significant improvements on both transmission estimation and haze removal results.

Vibration measurement serves as the basis for various engineering practices such as natural frequency or resonant frequency estimation. As image acquisition devices become cheaper and faster, vibration measurement and frequency estimation through image sequence analysis continue to receive increasing attention. In the conventional photogrammetry and optical methods of frequency measurement, vibration signals are first extracted before implementing the vibration frequency analysis algorithm. In this work, we demonstrated that frequency prediction can be achieved using a single feed-forward convolutional neural network. The proposed method is verified using a vibration signal generator and excitation system, and the result obtained was compared with that of an industrial contact vibrometer in a real application. Our experimental results demonstrate that the proposed method can achieve acceptable prediction accuracy even in unfavorable field conditions.