Occluded pedestrian re-identification is a challenging task in the field of computer vision. A method was proposed using the FGMS-Net network, which significantly enhances pedestrian re-identification in occluded environments through several improvements. Firstly, an improved foreground segmentation technique was employed to effectively remove background and other clutter information, resulting in more accurate feature extraction. Secondly, to address the occlusion issue, a multi-scale feature discrimination method was introduced, enabling the model to better capture local features and thereby enhancing identification capability. Finally, an attention mechanism was added to the backbone network, allowing the network to focus more on critical information and improve overall recognition performance. The experimental results show that method proposed has achieved significant performance improvement in the task of pedestrian re recognition with occlusion. On the Occluded-DukeMTMC dataset, the cumulative matching feature Rank-1 and mean average precision (mAP) reach 71.7% and 61.6%, respectively.

Although the multi-task convolutional neural networks (MTCNN) face detection algorithm has achieved good results in some face recognition tasks, the accuracy of face detection needs to be improved in the face of some complex small-scale and multi-person face detection tasks. An improved MTCNN algorithm was proposed. Firstly, the intersection over union (IoU) threshold parameter was fine-tuned when creating the data set to classify face samples more accurately. Secondly, replacing the max pooling layer of the network with convolutional layers can improve network performance. Finally, the squeeze-excitation(SE) attention mechanism was introduced into the O-Net network to improve the feature expression ability of the network. The test results show that compared with the original MTCNN algorithm, the detection accuracy of the P-Net network and R-Net network of the improved algorithm has increased by 1%, and the detection accuracy of the O-Net network has increased by 0.5%. Moreover, the improved algorithm performs better in the actual face detection task.

In dense scenes, the frequent occurrence of occluded or small-scale pedestrian objects poses significant challenges to traditional object detection models, frequently leading to a high number of missed detections and false positives. In order to solve the problem of high false negative rate and false positive rate in pedestrian detection in such dense scenes, a novel dense scene pedestrian detection framework called ST-YOLO was proposed. Firstly, the low-level small object detection layer in YOLOv5's backbone network was integrated into the feature pyramid network and path aggregation network structure, adding a pedestrian detection layer for detecting small objects. Secondly, the neck network of YOLOv5 was improved by utilizing multi-scale global information based on Swin Transformer and local information extracted by convolutional neural networks (CNN) to construct aggregated features and enhance the network's feature extraction capability. And the SIoU (scalable IoU) loss function was introduced in the prediction process to accelerate the convergence speed of the model and improve detection capability. Finally, Soft NMS (soft non maximum suppression) was used instead of the original non maximum suppression (NMS) algorithm to reduce the problem of mistakenly deleting detection boxes during the non maximum suppression stage and lower the false alarm rate of the detection algorithm. A large number of experiments on the Wide Person dataset have shown that the improved ST-YOLO algorithm has improved accuracy and mAP_0.5 by 5.7% and 3.6% respectively compared to the current mainstream YOLOv9 algorithm.

With the expansion of the Internet scale and changes in its topological structure, network management is facing huge challenges. Segment routing (SR) protocols, especially SRv6(segment routing over IPv6), have become a research hotspot due to their high programmability and scalability. The path optimization control mechanism based on SRv6 solves the problem of avoiding and relaying specific nodes in multiple demands and scenarios to improve network performance. A path transfer scheme for fully deployed SRv6 networks was proposed, and the routing overhead was reduced through the optimization of forked paths. For some deployment networks, define the critical path and design the avoidance and relay path forwarding scheme to optimize the path forwarding efficiency. The experimental results show that when SRv6 is fully deployed, the optimization scheme can effectively reduce the depth of the segment list and the routing overhead. In some deployment networks, only a small number of SRv6 nodes can achieve performance close to that of a full SRv6 network, successfully solving the problem of evading and relaying specific nodes. The research results provide effective support for the application of SRv6 in different network deployments.

The previous structural seismic vulnerability analysis is generally based on the characteristics of the structure itself, it is rare to combine with the differences of engineering sites in the study area, the location differences of different engineering sites within the city are ignored. Taking Chengdu City as the research area and the three-story reinforced concrete frame structure as an example, an analysis method for the seismic vulnerability of reinforced concrete frame structures based on peak ground acceleration (PGA) and the maximum inter-story displacement angle θ_max of the structure was proposed. For three-story reinforced concrete frame structures, this method conducts dynamic time-history analysis using the interlayer shear model to obtain θ_max under each seismic response. Then, logarithmic linear fitting is performed on θ_max and its corresponding ground motion to obtain the relationship between the two. For the Chengdu area, this method takes the historical ground motion data of Chengdu as the data basis and combines PGA calculation formula to obtain the PGA of each engineering site location in Chengdu. Furthermore, taking the maximum inter-layer displacement angle as the structural damage index and PGA as the ground motion intensity index, the highest structural failure probabilities of the structure under four different performance levels of full operation, basic operation, life safety and near collapse were studied, which were 94.1%, 89.1%, 74.7% and 40.8% respectively. Moreover, the overall changing trend of the structural failure probability at each performance level of the structure decreases from the west to the east. Therefore, the seismic construction requirements for structures in the western region can be appropriately strengthened, and those for structures in the eastern region can be appropriately relaxed, so as to save economic costs. The proposed method has certain application value in reducing the losses caused by earthquakes and provides a certain theoretical basis for the seismic design of building structures.

With the implementation of river dredging projects in China, a large amount of dredged silt has been generated. The treatment and disposal of silt have gradually attracted great attention. Using solidification technology is one of the effective ways to solve the problems caused by dredged silt. Taking the dredged silt from Beibaidang in Zhejiang as the research object, the solidified products were analyzed through X-ray diffraction (XRD), and the porosity and pore structure of the solidified silt soil were quantitatively analyzed by means of X-ray computed tomography (X-CT) and mercury intrusion porosimetry (MIP) tests. Meanwhile, it explores the mechanical variation laws and solidification mechanism of the soil under soaking and dry-wet cycling conditions. The research shows that the calcite content in the solidified soil increases with the increase of the solidifying agent dosage. The increase of ordinary sand dosage improves the small and medium-sized porosity inside the soil, but the overall porosity shows a decreasing trend. The results of water stability tests indicate that the strength and stability of the solidified silt soil are significantly improved with the increase of the solidifying agent content, and they first increase and then decrease with the increase of sand dosage.

In order to explore the influence of pore defects on the mechanical properties of concrete road and seek an equivalent model to replace porous concrete road to reduce computational time. Based on micromechanics methods, the effective elastic modulus, Poisson's ratio, coefficient of thermal conductivity and coefficient of thermal expansion of porous concrete were calculated. Three-dimensional double-layer concrete roads with randomly distributed, non-interference and varying sizes of spherical pores and their equivalent models were established to study their mechanical properties under three working conditions, namely, concentrated force, static vehicle load, and temperature-static vehicle load coupling, and further the simulation calculation time for each model was compared. The results show that under the coupling effect of temperature and static vehicle load, the increase of porosity has little effect on the temperature and displacement of porous concrete roads at the same depth and different times. Moreover, for the same porosity, the farther away from the pavement, the peak temperature shifts backward over time. When the porosity is within 8%, the actual porous model can be replaced by the Eshelby equivalent model, Mori-Tanaka equivalent model, or Self-Consistent equivalent model under concentrated force or static vehicle load, and by the equivalent model 2 under temperature-static vehicle load. With fixed computational power and constant porosity, the simulation time for the actual porous model far exceeds its equivalent model. Using equivalent models for research can significantly shorten the calculation time, and the computational efficiency can be approximately improved by about 99.8%.

Under the guidance of the full life cycle design concept, a scheme of moving the existing downspout pipe scheme to the reinforced concrete peripheral column was proposed. The main feature of this scheme is that the downspouts are pre-embedded into the surrounding columns and used for roof drainage during the building's service life. At the end of the building's lifespan, the downspouts were used to remove blast holes. To demonstrate the feasibility of this scheme, first of all, it was verified that the material and diameter of the embedded downspout could simultaneously meet the drainage function and the function of removing the blast hole. Secondly, by drawing on the axial embedded hole blasting technology of reinforced concrete beams and using the new type of blasting equipment, the long bag for loading explosives, the feasibility of using downspouts for blasting demolition was demonstrated from both the charge structure and the blasting operation aspects. Finally, the finite element analysis software ABAQUS was used to analyze the stress distribution and stress-strain laws when two types of cross-sectional peripheral columns were placed in four downspouts with different diameters. The analysis results show that when the hollow rate is small, the embedded pipe has little influence on the specimens. In order to avoid significant impact on the peripheral columns, the hollow rate of the specimens should not be greater than 2%. In combination with the requirements of blasting demolition, the hollow ratio should not be less than 0.18% either. Subsequently, it was pointed out that the relay service duration of pipes and concrete holes is sufficient to reach the building's life cycle. Finally, it is clarified that after the downspout is built in, it can also enhance the aesthetic appeal of the building and avoid the safety risk of thieves climbing along the downspout. Therefore, it is feasible to insert reinforced concrete peripheral columns into the downspout pipe.

To investigate the annual variation characteristics of soil temperature around the medium-shallow coaxial tube ground heat exchanger, a two-dimensional unsteady heat transfer model for the medium-shallow coaxial tube ground heat exchanger was established. The model was solved based on the finite volume method and validated using experimental data from the project. The research findings indicate that under individual heating and cooling conditions, the circulation mode of outer-in and inner-out has a greater impact on soil temperature at a depth of 100 m, while the inner-in and outer-out mode has a more significant effect on soil temperature at a depth of 500 m. During summer conditions, reverse heat transfer is more likely to occur with the outer-in and inner-out circulation mode. At the end of the first winter (or summer) season, the thermal influence radius of the soil is less than 10 m, but this radius increases over time. At the end of one operational cycle, the soil temperature increases at depths shallower than approximately 300 m and decreases at depths deeper than approximately 300 m. Orthogonal experiments reveal that the inlet water temperature in both winter and summer has a notable impact on temperature fluctuations at a soil depth of 100 m, while the inlet water temperature in summer significantly affects temperature fluctuations at a soil depth of 500 m.

Storm rainfall patterns are critical to infrastructure flood control, but the main design methods are hindered by requiring high resolution data and ignoring the impacts of climate change. Taking the typical small and medium-sized river basin (Liudong River Basin) with relatively high flood control pressure in the Karst area of southwest China over the years as an example, based on the Gamma distribution with flexible distribution characteristics and wide application, and using the duration data of the maximum 24-hour rainstorm in the past 60 years, the model performance was evaluated, the parameters of the model were determined, and the evolution characteristics of the parameters were analyzed. Based on the above results, the changing trends of rain pattern uncertainty, complexity and predictability were calculated. The results show that the model has high accuracy, and the average correlation coefficient is greater than 0.92. The evolution trend of the model parameters indicates that under the background of climate change, the rain pattern of the maximum 24-hour rainstorm shows the characteristics of decreasing shape factor and increasing scale factor. The uncertainty of rain patterns has generally increased, and the complexity and predictability show high spatial heterogeneity. The research results will provide references for optimizing the rainstorm rain pattern model and analyzing its dynamic evolution, thereby better preventing and controlling flood disasters, especially in Karst areas with high precipitation uncertainty and complex underlying surface conditions.