To promote the progress of fault diagnosis system, this study proposes an intelligent and effective fault diagnosis algorithm based on data-driven. Firstly, we propose a new time-frequency analysis method named second-order time-reassigned multisynchrosqueezing transform (STMSST) based on Gaussian-modulated linear group delay (GLGD) model to deal with the vibration signals of fault object for obtaining time-frequency images with high resolution. Then, an improved training method named evenly mini-batch training method is combined with convolutional neural network (CNN) to train and learn fault features from those obtained time-frequency images. Further, the proposed fault diagnosis algorithm is tested on the Case Western Reserve University (CWRU) bearing dataset and the Machinery Failure Prevention Technology (MFPT) Society dataset, respectively, and the experimental results indicate that the feature representation and training effect in our method is superior to state-of-the-art fault diagnosis methods. Finally, the proposed method is applied on the loudspeaker pure-tone detection dataset, which achieves the loudspeaker anomaly diagnosis, and the diagnosis result has verified that the method can meet the needs of practical engineering.

In this paper, a novel method, to the best of our knowledge, of structured light fields based on point cloud adaptive repair is proposed to realize 3D reconstruction for highly reflective surfaces. We have designed and built a focused light field camera whose spatial and angular resolution can be flexibly adjusted as required. Then the subaperture image extraction algorithm based on image mosaic is deduced and presented to obtain multidirectional images. After that, the 3D reconstruction of structured light field imaging based on point cloud adaptive repair is presented to accurately reconstruct for highly reflective surfaces. In addition, a method based on smoothness and repair rate is also proposed to objectively evaluate the performance of the 3D reconstruction. Experimental results demonstrate the validity of the proposed method to perform high-quality depth reconstruction for highly reflective surfaces. Generally, our method takes advantage of the multidirectional imaging of the light field camera and can ensure good modulation effect of structured light while avoiding hardware complexity, which makes it application more convenient.

The existing ultrasonic thickness measurement systems require high sampling frequencies for echo signal acquisition, leading to complex circuit designs and high costs. Moreover, extracting the characteristics of ultrasonic echo signals for accurate thickness measurement poses significant challenges. To address these issues, this paper proposes a method that utilizes conventional sampling frequencies to acquire high-frequency ultrasonic echo signals, overcoming the limitations of high-frequency data acquisition imposed by the Nyquist–Shannon sampling theorem. By employing an improved sampling reconstruction technique, the multi-cycle sampling signals are reconstructed and rearranged within a single cycle, effectively increasing the equivalent sampling frequency. Additionally, a combination of coarse estimation using fast Fourier transform (FFT) and precise phase extraction using the moving sine fitting algorithm is proposed for accurate thickness measurement, resolving the limitations of common thickness measurement methods such as peak detection, envelope detection, and Hilbert autocorrelation in terms of low measurement accuracy. Experimental results obtained from thickness measurements on 45 steel ultrasonic test blocks within the range of 3 mm to 20 mm indicate a measurement error of ±0.01 mm, while for thicknesses ranging from 1 mm to 50 mm, the measurement error is ±0.05 mm.