The detection of surface defects in concrete bridges using deep learning is of significant importance for reducing operational risks, saving maintenance costs, and driving the intelligent transformation of bridge defect detection. In contrast to the subjective and inefficient manual visual inspection, deep learning-based algorithms for concrete defect detection exhibit remarkable advantages, emerging as a focal point in recent research. This paper comprehensively analyzes the research progress of deep learning algorithms in the field of surface defect detection in concrete bridges in recent years. It introduces the early detection methods for surface defects in concrete bridges and the development of deep learning. Subsequently, it provides an overview of deep learning-based concrete bridge surface defect detection research from three aspects: image classification, object detection, and semantic segmentation. The paper summarizes the strengths and weaknesses of existing methods and the challenges they face. Additionally, it analyzes and prospects the development trends of surface defect detection in concrete bridges.

An improved cycle-consistent generative adversarial network (CycleGAN) method for defect data augmentation based on feature fusion and self attention residual module is proposed to address the insufficiency of defect sample data for light guide plate (LGP) in production, as well as the problem of minor defects. Two optimizations are made to the generator of CycleGAN: fusion of low resolution features obtained from partial up-sampling and down-sampling with high-resolution features, combination of self attention mechanism with residual network structure to replace the original residual module. Qualitative and quantitative experiments were conducted to compare different data augmentation methods, and the results show that the defect images of the LGP generated by the improved network were more realistic, and the accuracy of the you only look once version 5 (YOLOv5) detection network for the LGP was improved by 5.6%, proving the effectiveness and accuracy of the proposed method.

In this paper, a two-stage light detection and ranging (LiDAR) three-dimensional (3D) object detection framework is presented, namely point-voxel dual transformer (PV-DT3D), which is a transformer-based method. In the proposed PV-DT3D, point-voxel fusion features are used for proposal refinement. Specifically, keypoints are sampled from entire point cloud scene and used to encode representative scene features via a proposal-aware voxel set abstraction module. Subsequently, following the generation of proposals by the region proposal networks (RPN), the internal encoded keypoints are fed into the dual transformer encoder-decoder architecture. In 3D object detection, the proposed PV-DT3D takes advantage of both point-wise transformer and channel-wise architecture to capture contextual information from the spatial and channel dimensions. Experiments conducted on the highly competitive KITTI 3D car detection leaderboard show that the PV-DT3D achieves superior detection accuracy among state-of-the-art point-voxel-based methods.

Manufacturing and agricultural industries use manual methods to count materials. This leads to low accuracy and inefficiency. This paper proposes a secondary counting method that combines main and differential counting. The area-fill identification algorithm is applied to mark the counted materials. To verify the effectiveness of the proposed counting algorithm, numbers of countings are conducted for different materials, such as the screws, hole gaskets, beans, jujube, etc. The results show that the counting accuracy reaches 98% for materials with size of 2—20 mm. The method has delivered a high-efficiency and high-accuracy automatic intelligent counting, with a wide range of application prospects and reference value.

Most binary networks apply full precision convolution at the first layer. Changing the first layer to the binary convolution will result in a significant loss of accuracy. In this paper, we propose a new approach to solve this problem by widening the data channel to reduce the information loss of the first convolutional input through the sign function. In addition, widening the channel increases the computation of the first convolution layer, and the problem is solved by using group convolution. The experimental results show that the accuracy of applying this paper's method to state-of-the-art (SOTA) binarization method is significantly improved, proving that this paper's method is effective and feasible.

The bistratified lobula giant type 1 (BLG1) neuron is an identified looming-sensitive neuron in crab’s visual brain that demonstrates special sensitivity to diving targets, or descending approaching motions. In this paper, a novel neural model is proposed to shape such unique selectivity through incorporating a bio-plausible feedforward contrast inhibition synapse and a radially extending spatial enhancement distribution. Herein the synaptic connections and neuronal functions of this model are placed within a framework for matching and describing underlying biological findings. The systematic and comparative experiments have validated the proposed computational model that reconciles with the characteristics of BLG1 neurons in crab.

Multi-object tracking (MOT) techniques have been increasingly applied in a diverse range of tasks. Unmanned aerial vehicle (UAV) is one of its typical application scenarios. Due to the scene complexity and the low resolution of moving targets in UAV applications, it is difficult to extract target features and identify them. In order to solve this problem, we propose a new re-identification (re-ID) network to extract association features for tracking in the association stage. Moreover, in order to reduce the complexity of detection model, we perform the lightweight optimization for it. Experimental results show that the proposed re-ID network can effectively reduce the number of identity switches, and surpass current state-of-the-art algorithms. In the meantime, the optimized detector can increase the speed by 27% owing to its lightweight design, which enables it to further meet the requirements of UAV tracking tasks.

Implicit polynomials (IPs) are considered as a powerful tool for object curve fitting tasks due to their simplicity and fewer parameters. The traditional linear methods, such as 3L, MinVar, and MinMax, often achieve good performances in fitting simple objects, but usually work poorly or even fail to obtain closed curves of complex object contours. To handle the complex fitting issues, taking the advantages of deep neural networks, we designed a neural network model continuity-sparsity constrained network (CSC-Net) with encoder and decoder structure to learn the coefficients of IPs. Further, the continuity constraint is added to ensure the obtained curves are closed, and the sparseness constraint is added to reduce the spurious zero sets of the fitted curves. The experimental results show that better performances have been obtained on both simple and complex object fitting tasks.

Aero-optic imaging is a kind of optical effect, which describes the imaging deviation on the imaging plane. In this paper, the effect of the change of Mach number of blunt aircraft on the aero-optic imaging deviation is studied. The imaging deviations of Mach number 0.5-3 are analyzed systematically. The results show that with the increase of Mach number, imaging deviation increases gradually, and the increase rate is gradually slow. Imaging deviation slope decreases gradually with the increase of Mach number, and gradually tends to be zero, suggesting that imaging deviation is not sensitive to the change of the larger Mach number. In other words, the Mach number of smaller changes can lead to larger imaging deviation. As the Mach number of the aircraft increases, the slope of the imaging offset tends to be closer and closer to 0. When the Mach number of the aircraft increases to a certain extent, the change of the imaging offset will not have much influence. Therefore, in order to reduce the impact of flight speed on imaging migration, the aircraft should fly at a higher Mach number.

The resolution and quality of the depth map captured by depth cameras are limited due to sensor hardware limitations, which becomes a roadblock for further computer vision applications. In order to solve this problem, we propose a new method to enhance low-resolution depth maps using high-resolution color images. The structural-aware term is intro-duced because of the availability of structural information in color images and the assumption of identical structural features within local neighborhoods of color images and depth images captured from the same scene. We integrate the structural-aware term with color similarity and depth similarity within local neighborhoods to design a local weighting filter based on structural features. To use non-local self-similarity of images, the local weighting filter is combined with the concept of non-local means, and then a non-local weighting filter based on structural features is designed. Some ex-perimental results show that super-resolution depth image can be reconstructed well by the process of the non-local fil-ter and the local filter based on structural features. The proposed method can reconstruct much better high-resolution depth images compared with previously reported methods.

This paper designs an energy-quality multilevel framework for the coding and transmission of aerial images, and then introduces a scaling-based intra encoder with flexible sampling factor (SF) and quantization parameter (QP). By experimentally investigating how different coding configurations affect the complexity-rate-quality characteristics of aerial images, this paper derives a configuration estimation model between energy-quality level and appropriate (SF, QP) configuration. By utilizing the model, a bivariate control scheme is proposed so as to progressively adjust sender's energy consumption under quality constraints. The experimental results show that the proposed scheme can achieve better energy-quality tradeoff with a wider quality range, and reduce the energy consumption above a certain quality.

Dimensionality reduction is becoming an important problem in hyperspectral image classification. Band selection as an effective dimensionality reduction method has attracted more research interests. In this paper, a band selection method for hyperspectral remote sensing images based on subspace partition and particle frog leaping optimization algorithm is proposed. Three new evolution strategies are designed to form a probabilistic network extension structure to avoid local convergence. At the same time, the information entropy of the selected band subset is used as the weight of inter-class separability, and a new band selection criterion function is constructed. The simulation results show that the proposed algorithm has certain advantages over the existing similar algorithms in terms of classification accuracy and running time.

In order to get a satisfactory image fusion effect, getting a focus map is very necessary and usually difficult to finish. In this paper, we address this problem with a half weighted gradient approach, aiming to obtain a direct mapping be-tween focus map and source images. Based on the advantages of multi-scale weighted gradient, while abandoning the shortcomings of weighted gradient, a new multi-focus image fusion method called half weighted gradient and self-similarity (HWGSS) is proposed. Experimental results validate that the proposed algorithm can obtain state-of-the-art fusion performance in terms of both qualitative and quantitative evaluations.

Quantification of fiber orientation is the key to characterizing the tissue mechanical properties and diagnosing diseases. A center line-based algorithm is presented for estimating the orientation distribution that first skeletonizes a binary image of fibers, followed by orientation estimation using a weight vector summation algorithm along the center line of image. Then we use the orientation at the skeleton to approximate the orientation of each pixel between the boundary and skeleton. The algorithm is applied for characterizing collagen fibers of mouse skins in second harmonic generation (SHG) image, and the circle standard deviation of orientation could be a biomarker to differentiate the naturally aging skins.

This paper presents an image denoising method based on bilateral filtering and non-local means. The non-local region texture or structure of the image has the characteristics of repetition, which can be used to effectively preserve the edge and detail of the image. And compared with classical methods, bilateral filtering method has a better performance in denosing for the reason that the weight includes the geometric closeness factor and the intensity similarity factor. We combine the geometric closeness factor with the weight of non-local means, and construct a new weight. Experimental results show that the modified algorithm can achieve better performance. And it can protect the image detail and structure information better.

With the development ofthree-dimensional (3D) technology, visual fatigue problems in 3D video have got more attention. In this paper, we combine the human vision characteristics and depth perception theory, and propose a 3D video visual comfort evaluation method on the consistency of accommodation and convergence, which evaluates the visual comfort from the quantitative perspective under different horizontal disparities and viewing distances. The experimental results show that the proposed evaluation method exhibits good consistency with the subjective assessment results.