Aiming at the issue of how to realize the dynamic updating of the digital twin model of historic buildings, a “build-compare-update” updating idea was proposed. Firstly, the shape distribution method was used to draw the shape distribution curve of the digital twin model of the historical building at different time points, based on completing the construction of the digital twin model of the historical building. Secondly, the mean absolute error method was used to assess the similarity of the two models. Furthermore, according to the results of the similarity assessment, the updating strategy for the model of the components of the historic building was set up and the updating of the digital twin model was completed with the help of relevant software. Finally, to validate the design, the arch column component of the Jade Buddha Hall of Hongci Temple was used as a case study. The results show that the proposed idea of updating the historical building model and its implementation method not only lay the foundation for realizing the dynamic updating of the digital twin model of historical buildings, but also provide reference value for the intelligent protection of historical buildings.

Modeling social networks using directed graphs and calculating the influence of nodes in settlement graphs using deep learning methods are important branches in the field of social research. In order to solve the problem that graph neural networks cannot well calculate the influence of nodes based on constructed node features, inspired by GraphGPS, a heat module was designed to propose a method of calculating the influence of nodes that integrated semantic, behavioral, and heat information in real networks. Firstly, the self-information obtained based on nodes' multiple centrality and orthogonal distribution sampling was used as the initial semantic features of the nodes. Secondly, the node features were fused by graph neural network. Once again, the node heat information was learned by the heat module. Finally, the fusion of the extracted semantic, behavioral, and heat features was implemented to calculate the node influence. Experiments were conducted on four real network datasets. The results show that the model with the addition of the heat module can effectively calculate node influence.

The traditional steady-state visual evoked potential (SSVEP) brain computer interface system usually uses a small number of frequencies for encoding, resulting in a limited number of encodings to dozens, which cannot meet the demands of environmental tasks with a large number of instructions. To address this issue, a Hamming distance multi frequency code (HDMFC) paradigm and corresponding recognition algorithm based on Hamming distance were proposed. The Hamming distance was combined with stimulus paradigm encoding and signal recognition algorithms to encode 120 instructions using 8 frequency signals. Data collection and classification experiments were conducted on 7 subjects. The results show that the accuracy of the 120 encoding online experiment based on Hamming distance can reach 90.60%. The research results provide an effective method for increasing the number of SSVEP paradigm codes and improving classification performance, verifying the practicality and effectiveness of Hamming distance in this field.

In response to the challenges of low efficiency, high cost, and difficulty in deployment on mobile devices in current road damage detection technology, a novel road crack detection method based on the improved YOLOv8 algorithm, named YOLOv8 road crack (YOLOv8-RC), was proposed. The C2f module, based on the YOLOv8n architecture, was enhanced through the introduction of dynamic snake convolution technology, which more accurately identified tubular structural features and adaptively focuses on fine and curved local structures. Furthermore, a highly efficient multi-scale attention(EMA) mechanism was incorporated into the algorithm, effectively enhancing recognition accuracy. In the neck structure of the model, a weighted bidirectional pyramid network(BiFPN) was added to achieve multi-scale fusion of features, thereby optimizing both the accuracy and efficiency of the algorithm. Experimental results on the RDD2022-China-MotorBike and RDD2022-Japan datasets demonstrate that the improved YOLOv8n-RC model achieves mAP50 scores of 78.8% and 43.8%, respectively, representing improvements of 3.9% and 3% over the original YOLOv8n model. The total number of model parameters for the proposed algorithm is only 2.84 M, and the computational complexity is 7.8 G, underscoring the practicality and superiority of this method.

An improved DGA-YOLOv8 offshore ship target detection algorithm was proposed to tackle the issues of low accuracy and single ship detection categories that are present in traditional ship target detection algorithms. Firstly, the network was adapted to include deformable convolution, which expanded the model's receptive field. Learnable offsets were introduced, allowing the model to adaptively adjust the size and shape of the receptive field in response to the actual shape of the object, ensuring that the convolution area can precisely cover the contour of the ship object. Secondly, the incorporation of a GAM(global attention mechanism) attention mechanism enabled the network to effectively emphasize the key features of ship targets, thereby enhancing the target recognition capability. The experimental results demonstrate that the improved algorithm achieves accuracy and average accuracy mean (mAP) of 96.4% and 92.2%, respectively. An frames per second(FPS) of 43.55 is recorded, indicating not only an enhancement in accuracy but also the maintenance of a certain detection speed, thus fulfilling the requirements for real-time detection. When compared with other mainstream algorithms, such as faster region-based convolutional neural network(Faster R-CNN) and YOLOv5s, YOLOv10. The results show that the proposed algorithm exhibits higher average accuracy and significant superior classification performance.

In order to improve the accuracy of in-camera parameter calibration and reduce the error caused by inaccurate estimation of corner coordinates, a subpixel iterative corner optimization algorithm was proposed. By modeling the camera image, integrating the distortion mathematical model, and using the subpixel iterative algorithm, the gradient value changes of points in the search window were calculated and the corner coordinates were iteratively optimized to provide more accurate initial values for calibration. Combining with the optimized corner coordinates, Zhang's calibration method was used to solve the internal parameters, and the influence of corner points on the acquisition of diagonal points and camera calibration was discussed through experiments. The calibration accuracy was characterized by reprojection error, and the effectiveness of the iterative algorithm was tested. The experimental results show that the algorithm can effectively improve the accuracy of corner acquisition and the calibration accuracy of camera parameters by using fewer iterations.

Remote sensing images are characterized by diverse scales, dense arrangement and small target sizes, etc. Aiming at the problem that there is much background noise in remote sensing images and vehicle targets are small and difficult to be acquired. A vehicle target detection algorithm based on improved feature fusion method, Atiny-YOLO was proposed. Firstly, an additional detection layer for small targets was introduced into the Neck layer of YOLOv5 so as to generated a small target detection algorithm for drone remote sensing images. Neck layer to introduce an additional detection layer for small targets, so as to generated a larger-scale feature map and effectively identified the detailed features of small objects. Secondly, a split operation was added to the C3 module to reuse the image feature information, and the Swin Transformer module was further optimized to improve the usage rate of the effective information. Lastly, by improving the feature fusion channel, the detection accuracy was improved while the model parameters were reducing the model parameters. The Atiny-YOLO algorithm was tested on the AU-AIR(aerial universal autonomous inspection and recognition) dataset. The experimental results show that the average detection accuracy of the Atiny-YOLO algorithm compared to the baseline algorithm is improved by about 2.9%. It reaches 95.5% and the detection speed reaches 234 frames/s. These results verify that the Atiny-YOLO algorithm meets the real-time performance while the model detection accuracy is greatly improved.

In order to study the mechanical properties of the column foot joint of a self-resetting cylindrical pier, the mechanical mechanism of self-resetting cylindrical pier column foot joint with circular section was analyzed theoretically. The calculation formula of key points of the whole force and displacement hysteresis curve of the pier was derived, including yield point, failure point, etc. The theoretical analysis model of the relationship between pier jacking force and displacement was established. Based on OpenSees platform, fiber hinge model was used to model the section of pillar foot of self-resetting energy-consuming pier. Combined with pseudo-static test results, the feasibility and accuracy of the fiber hinge model were verified. On this basis, a self-resetting cylindrical pier was set up by fiber hinge model, and the key point of the deduced force-displacement hysteresis curve was compared with the skeleton curve of the simulated cylindrical pier to verify the accuracy of the deduced method. The results show that the derivation method can obtain the pressure relief point, yield point and failure point of self-resetting cylindrical pier, which can provide reference for the research of self-resetting pier with circular section in the future.

In order to study the applicability of the assembled subway station structure in the liquefiable site of the seismic fortification area, based on the engineering practice of Shuangfeng Station of Changchun Metro, the seismic response analysis of the assembled subway station structure in the liquefiable stratum was carried out by using the finite difference software FLAC^3D. Compared with the calculation results of the corresponding cast-in-place subway station conditions, the pore water pressure of the foundation, the dynamic response and floating characteristics of the subway station structure and the deformation characteristics of the assembled subway station structure were studied. The results show that, similar to the calculation condition of cast-in-place underground structure, in the calculation condition of assembled subway station in liquefiable site, the pore pressure ratio of site soil shows three stages: initial slow increase stage, middle rapid increase stage and final gentle stage. When the seismic intensity is high, the underground structure has a significant inhibitory effect on the liquefaction of the surrounding soil. The acceleration response of the prefabricated station structure is smaller than that of the cast-in-place structure, which reflects its flexible adaptation to the deformation of the terrain in the liquefiable site. The connection between the side wall and the middle plate is the key part of structural strength control. When the intensity of the input ground motion is low, the vertical displacement changes of the structure are manifested at different positions of the bottom plate show the development stage of a small amount of sinking at the beginning, then a sharp rise, and finally a downward trend, but with the increase of the intensity of the ground motion, this phenomenon is gradually not obvious. The deformation of the assembled subway station structure under horizontal ground motion is closer to the shear deformation. Its flexible connection is easier to adapt to the larger vertical relative deformation than the rigid connection of the cast-in-place station, and its anti-overturning ability is better.

In order to realize the optimal design of rock-socketed pile foundation in mountainous area and effectively prevent the rupture of the rock, the compression mechanical behaviors and failure mechanisms of brittle rocks with different lithologies were experimentally and numerically investigated. Firstly, uniaxial compression tests were carried out on five kinds of brittle rocks based on the self-made compression testing system. Acoustic emission (AE) monitor system and digital image correlation (DIC) technology were used to discuss the characteristic parameters of AE signals during the crushing. Moreover, the characteristics of critical failure precursors and the evolution rules of damage failure of different rocks were obtained by using the nonlinear explicit finite element (FE) method. The experimental results were in good agreement with the corresponding FE ones. The results show that there are the obvious AE phenomena in brittle rocks during the uniaxial compression test, and the active degree of ringing count and accumulated energy can describe the internal damage of the specimens effectively. At the initial stage of the compression, there is no obvious crack on the rock surface. However, more cracks appear on the surface of the specimens with the increase of the compressive stress. The specimens are destroyed instantaneously when the compressive stress reaches the peak stress, and no new cracks are generated. The distribution and evolution of local strains can intuitively predict the breeding and expansion of new fissures. The research results provide theoretical reference for monitoring and warning of rock failure and foundation design of transmission line in mountainous area.