Xiao Liang

Identifying wheelset tread defects for heavy-haul trains is an increasingly valued yet challenging task. These defects are usually characterized by irregular shapes, wide coverage, and high codependency under prolonged heavy loads. Frequently, multiple correlated defects are observed to coexist and intertwine within the same field of view, posing substantial difficulties for fine-grained identification. Although recent methods have progressed much, they neglect inter-label dependencies and cannot distinguish well between coexisting tread defects. To this end, we explore for the first time the importance of inter-label dependencies in identifying coexisting tread defects of heavy-haul trains. Concretely, we propose a novel multi-label classification framework based on a dual-branch united image-graph network to effectively recognize the three most concerning coexisting defects on heavy-haul train wheelset treads: crack, exfoliation, and spalling. To the best of our knowledge, this is the first work to consider multi-label learning for fine-grained tread defect identification. In the image-oriented branch, we design a pyramid feature re-extractor and a semantic attention guider to excavate expressive feature representation and ensure effective multi-scale fusion. In the graph-oriented branch, we construct graph convolutional networks to exploit sufficient label dependency priors. Further, a multi-level merging mechanism between label embeddings and image features is proposed to enhance the complementation between image and graph domains. Experiment results demonstrate superior performance than other competitive methods in identifying coexisting tread defects for heavy-haul trains. Additionally, the proposed method shows good generalization ability in similar industrial inspection scenes for rail surface defects and aluminum profile defects.

The detection of surface weak scratches is an intractable but vital task in optics-centered industries. However, the intrinsic characteristics of weak scratches, such as a narrow width, long span, and shallow depth, make it extremely difficult to effectively detect these scratches. In this paper, we tackle this issue from two perspectives. First, a multimodal microscopic imaging system is created by combining discrete multispectral illumination with linear polarization. Imaging experiments demonstrated that this system could highlight more scratch details, improve image clarity, and alleviate the image blur problem induced by wide spectrum scattered lights. Second, a scratch-oriented U-shaped deep encoder–decoder network equipped with optimized residual encoding modules, serial–parallel multiscale fusion modules, and triple-convolution decoding modules is proposed to segment the weak scratches from a raw image. The detection experiments demonstrate that our model can accurately segment the weak scratches on optical surfaces and achieve better detection performance using significantly fewer parameters compared to similar deep learning models. Meanwhile, experiments on the building crack dataset prove the excellent generalization capability.

Blur detection is aimed to recognize the blurry pixels from a given image, which is increasingly valued in vision-centered applications. Albeit great improvement achieved by recent deep learning-based methods, the overweight model and rough boundary still pose challenges to blur detection. In this paper, we propose a Hierarchical Edge-guided Region-complemented Network (HER-Net) to tackle the above issues in quest of a favorable accuracy–complexity trade-off. First, we propose novel olive-shaped and pear-shaped inverted bottleneck structures based on large-kernel depth-wise convolutions to build a very concise architecture. Second, we provoke and exploit region-concerned and edge-concerned morphological priors to refine the boundary. To this end, we propose a reverse-region spatial attention to mine the complementary affinities between blurry and sharp regions so as to enrich the residual details around the boundary. In addition, we propose an edge spatial attention to guide the edge-concerned cues to emphasize the features related to the boundary. Both attentions are embedded into the model with hierarchical manners. Extensive experiments on three benchmark datasets demonstrate that the proposed method can achieve better detection performance using fewer parameters and lower floating-point operations compared to competitive methods. It proves the efficiency and effectiveness of our method in blur detection task.

Surface defect identification plays a vital role in defective component rapid screening tasks in optics-related industries. However, the weakness and complexity of optical surface defects pose considerable challenges to their effective identification. To this end, a deep network based on multi-scale mixed kernels and structural re-parameterization is proposed to identify four manufacturing and two non-manufacturing optical surface defects. First, we design a multi-size mixed convolutional kernel with multiple receptive fields to extract rich shallow features for characterizing the defects with varying scales and irregular shapes. Then, we design an asymmetric mixed kernel integrating square, horizontal, vertical, and point convolutions to capture rotationally robust middle-and-deep features. Moreover, a structural re-parameterization strategy is introduced to equivalently convert the multi-branch architecture in the training phase into a deploy-friendly single-branch architecture in the inference phase, so that the model can obtain higher inference speed without losing any performance. Experiments on an optical surface defect dataset demonstrate that the proposed method is efficient and effective. It achieves a remarkable accuracy of 97.39% and an ultra-fast inference speed of 201.76 frames/second with only 5.23M parameters. Such a favorable accuracy–speed trade-off is capable of meeting the requirements of real-world optical surface defect identification applications.

The adaptive iterative learning control method for electro-hydraulic shaking tables based on the complex optimization algorithm was proposed to overcome the potential stability problem of the traditional iteration control method. The system identification precision’s influence on convergence was analyzed. Based on the real optimization theory and the mapping relationship between real vector space and complex vector space, the complex Broyden optimization iterative algorithm was proposed, and its stability and convergence was analyzed. To improve the stability and accelerate the convergence of the proposed algorithm, the complex steepest descent algorithm was proposed to cooperate with the complex Broyden optimization algorithm, which can adaptively optimize the complex steepest gradient iterative gain and update the system impedance in real time during the control process. The shaking tables experiment system was designed, applying xPC target rapid prototype control technology, and a series of experimental tests were performed. The results indicated that the proposed control method can quickly and stably converge to the optimal solution no matter whether the system identification error is small or large, and, thus, verified that validity and feasibility of the proposed adaptive iterative learning method.

Blur detection is aimed to differentiate the blurry and sharp regions from a given image. This task has attracted much attention in recent years due to its importance in computer vision with the integration of image processing and artificial intelligence. However, blur detection still suffers from problems such as the oversensitivity to image noise and the difficulty in cost–benefit balance. To deal with these issues, we propose an accurate and efficient blur detection method, which is concise in architecture and robust against noise. First, we develop a sequency spectrum-based blur metric to estimate the blurriness of each pixel by integrating a re-blur scheme and the Walsh transform. Meanwhile, to eliminate the noise interference, we propose an adaptive sequency spectrum truncation strategy by which we can obtain an accurate blur map even in noise-polluted cases. Finally, a multi-scale fusion segmentation framework is designed to extract the blur region based on the clustering-guided region growth. Experimental results on benchmark datasets demonstrate that the proposed method achieves state-of-the-art performance and the best balance between cost and benefit. It offers an average F₁ score of 0.887, MAE of 0.101, detecting time of 0.7 s, and training time of 0.5 s. Especially for noise-polluted blurry images, the proposed method achieves the F₁ score of 0.887 and MAE of 0.101, which significantly surpasses other competitive approaches. Our method yields a cost–benefit advantage and noise immunity that has great application prospect in complex sensing environment.

Segment assembling is one of the principle processes during tunnel construction using shield tunneling machines. The segment erector is a robotic manipulator powered by a hydraulic system to assemble prefabricated concrete segments onto the excavated tunnel surface. Nowadays, automation of the segment erector has become one of the definite developing trends to further improve the efficiency and safety during construction; thus, closed-loop motion control is an essential technology. Within the segment erector, the lifting gantry is driven by dual cylinders to lift heavy segments in the radial direction. Different from the dual-cylinder mechanism used in other machines such as forklifts, the lifting gantry usually works at an inclined angle, leading to unbalanced loads on the two sides. Although strong guide rails are applied to ensure synchronization, the gantry still occasionally suffers from chattering, “pull-and-drag”, or even being stuck in practice. Therefore, precise motion tracking control as well as high-level synchronization of the dual cylinders have become essential for the lifting gantry. In this study, a complete dynamics model of the dual-cylinder lifting gantry is constructed, considering the linear motion as well as the additional rotational motion of the crossbeam, which reveals the essence of poor synchronization. Then, a two-level synchronization control scheme is synthesized. The thrust allocation is designed to coordinate the dual cylinders and keep the rotational angle of the crossbeam within a small range. The motion tracking controller is designed based on the adaptive robust control theory to guarantee the linear motion tracking precision. The theoretical performance is analyzed with corresponding proof. Finally, comparative simulations are conducted and the results show that the proposed scheme achieves high-precision motion tracking performance and simultaneous high-level synchronization of dual cylinders under unbalanced loads.

Blur detection is aimed at estimating the probability of each pixel being blurred or non-blurred in an image affected by motion or defocus blur. This task has gained considerable attention due to its promising application fields in computer vision. Accurate differentiations of anomalous regions (including the sharp but homogeneous regions and pseudo-sharp backgrounds) and motion-blurred regions are main challenges in blur detection, in which both conventional and recently developed blur detection methods have limited performance and low time efficiency. To address these issues, this paper develops an accurate and fast blur detection method for both motion and defocus blur using a new end-to-end deep neural network. First, a novel multi-input multi-loss encoder-decoder network (M-shaped) is proposed to learn rich hierarchical representations related to blur. Then, to resolve the problem that blur degree is susceptible to scales, we construct a pyramid ensemble model (PM-Net) consisting of different scales of M-shaped subnets and a unified fusion layer. Experiments demonstrate that the proposed PM-Net can accurately handle those challenging scenarios with anomalous regions for both defocus and motion blur. Our method performs better than previous state-of-the-art methods. It achieves the F1-score of 0.893 for only defocus blur and 0.884 for joint motion and defocus blur, both of which significantly surpass previous methods on the benchmark BDD dataset. We also test our PM-Net on another public CDD dataset composed of challenging defocused images. The proposed method also outperforms other published methods with an F1-score of 0.885. In addition, our proposed method is hundreds of times faster (millisecond) than other state-of-the-art methods (second). Moreover, our experiments also demonstrate that the PM-Net is robust to noise and has a good generalization property.