Global inspection of large-scale tunnels is a fundamental yet challenging task to ensure the structural stability of tunnels and driving safety. Advanced LiDAR scanners, which sample tunnels into 3D point clouds, are making their debut in the Tunnel Deformation Inspection (TDI). However, the acquired raw point clouds inevitably possess noticeable occlusions, missing areas, and noise/outliers. Considering the tunnel as a geometrical sweeping feature, we propose an effective tunnel deformation inspection algorithm by extracting the global spatial axis from the poor-quality raw point cloud. Essentially, we convert tunnel axis extraction into an iterative fitting optimization problem. Specifically, given the scanned raw point cloud of a tunnel, the initial design axis is sampled to generate a series of normal planes within the corresponding Frenet frame, followed by intersecting those planes with the tunnel point cloud to yield a sequence of cross sections. By fitting cross sections with circles, the fitted circle centers are approximated with a B-Spline curve, which is considered as an updated axis. The procedure of “circle fitting and B-SPline approximation” repeats iteratively until convergency, that is, the distance of each fitted circle center to the current axis is smaller than a given threshold. By this means, the spatial axis of the tunnel can be accurately obtained. Subsequently, according to the practical mechanism of tunnel deformation, we design a segmentation approach to partition cross sections into meaningful pieces, based on which various inspection parameters can be automatically computed regarding to tunnel deformation. A variety of practical experiments have demonstrated the feasibility and effectiveness of our inspection method.

We present a novel correspondence-learning model for real-time registration of partially overlapping point clouds, between which the relative translation is large and hence identifying the correspondences is challenging. Our goal is to improve the feature learning for accurate correspondence establishment, which enables the promotion of registration performance significantly in terms of efficiency. This is realized by two particular designs. The first is a graph-based feature extraction module, which aggregates both inter and intra contexts of the input point clouds simultaneously for strengthening the connection between inputs. The second is a feature refinement module, which uses sparse convolutions to further widen the feature differences of dissimilar structures. The two modules reinforce each other to improve correspondence learning for robust and fast point cloud registration with low overlap. We evaluate the method on both synthetic and real-world large-scale datasets. The results in real registration tasks show that our method attains competitive registration accuracy with state-of-the-art methods, and is almost two times faster than these competing methods in some scenarios.

Point cloud denoising and normal estimation are two fundamental yet dependent problems in digital geometry processing. However, both are often independently researched, leading to inconsistent geometry on 3-D surfaces. To address it, we propose point-and-normal interactive network (PN-Internet), an end-to-end PN-Internet work for joint point cloud denoising and normal estimation. PN-Internet leverages the geometric dependency between point positions and normals to design two interactive graph convolutional networks (GCNs): a point-to-normal network and a normal-to-point network. It adopts a coarse-to-fine learning paradigm, where two GCNs are exploited to, respectively, perform point cloud denoising and normal estimation. The point-to-normal network improves the quality of the normals using an MLP module, while the normal-to-point network refines the point positions using a parameter-free projection module based on the constraints from the normals. In addition, we introduce a feature-aware loss function to preserve the quality of 3-D shape features. Unlike most existing methods, PN-Internet takes advantage of the geometric dependency between points and normals and benefits from training data. Our experimental results demonstrate that PN-Internet achieves geometric consistency between point cloud denoising and normal estimation. Furthermore, we show significant improvements over state-of-the-art methods.

The objective of 3D single object tracking is to determine the location of a target in a series of point clouds. However, this task still faces several challenges, such as the sparsity of the point cloud and the timeliness requirement of models. In this letter, we propose using an iterative siamese framework to improve tracking performance. We propose a feature learning backbone that captures correlations between different data samples. Next, we design an iterative siamese attention for feature enhancement of both the template and search to further distinguish the object from the background. Specifically, to perform the information interaction between the two, we employ a cross-attention operation on the template and the search, respectively. To enhance the correlations of the interior of the point cloud, we apply a self-attention operation on them. We repeat this cross-attention and self-attention operation several times to refine the point cloud feature learning. Following the iterative siamese attention module, we input the search feature into the Bird Eye View(BEV) target location network and utilize three prediction heads to predict the target bounding box. Experiments demonstrate that our method consistently outperforms several state-of-the-art approaches in some categories by a significant margin.

Woven composite fan blade preforms are essential structural components in modern aeroengines, where the quality of surface forming directly influences aerodynamic performance and structural reliability. Deep learning based reliable detection of surface defects is still very challenging, owing to the irregular shapes of defects in real production environments and the scarcity of data. To address these challenges, we propose a Yarn and Preform Surface Defect Detection System tailored for woven aeroengine fan blade preforms. The system integrates a YPSDD Image Acquisition Station with a specialized deep neural network YPSDD-Net. YPSDD-Net comprises a main detection branch and a few-shot adaptation branch. The main detection branch, built on a YOLOv8 backbone, integrates a Slender Mamba module and a Category-Aware Head (CAHead) to capture and localize complex defects. The Few-shot Adaptation branch enables rapid adaptation to novel defect types using limited labeled samples. We evaluate the proposed system on four datasets, including two in-house industrial datasets and two external benchmarks, to validate both accuracy and cross-domain generalization. Comparative experiments show that YPSDD-Net consistently outperforms state-of-the-art detectors across key metrics, achieving a precision of 96.2% on the Carbon Fiber Yarn Defects Dataset and 89.8% on the Woven Preform Surface Defects Dataset. Ablation studies and on-site assessments further confirm its robustness and suitability for real-time industrial deployment.

Based on the parallel classification of library borrowing data, the dynamic management and service ability of library borrowing information are improved, and a parallel classification method of library borrowing data based on Bayesian is proposed. According to the borrowing reports of user groups and individuals, a sampling model of library borrowing data is constructed, and a comprehensive borrowing index analysis model for users' personal construction system is established. Bayesian classification model is used to realize the feature mining and classification of borrowing data, and the correlation between different types of literature resources is mined according to the results of spectral feature extraction. Combined with the borrowing intensity between users with different attributes and different book classifications, Naive Bayesian classification algorithm is used to fully explore the internal relationship between fields. Based on the causal relationship between influencing factors, a dynamic model of parallel classification of library borrowing data is constructed, and the library borrowing data is matched according to certain rules. After weighted calculation, the reading behavior is quantified in the form of comprehensive index to realize parallel classification of library borrowing data. The test results show that the parallel classification of library borrowing data by this method improves the mining efficiency and accuracy of borrowing data, and the parallel processing efficiency is high.