The northern Guangdong-southern Hainan region is located at the intersection of Eurasian Plate, Pacific Plate, and Indian Plate. Since Late Mesozoic, this region has been influenced by multiple phases of tectonic activities. Its thermal evolution history has garnered widespread attention due to its ability to constrain regional tectonic evolution process. This study has collected, compiled, and processed published low-temperature thermochronological data from northern Guangdong-southern Hainan region using inverse distance weighting(IDW) interpolation method. By analyzing the age distribution and thermal history inversion modeling results, it restores its thermal evolution history since Late Mesozoic ( approximately 150 Ma) and explores the associated tectonic process. The thermal history of Hainan Island can be broadly divided into four stages: two episodes of very rapid cooling rate at 110-90 Ma and 40-20 Ma, interspersed with two periods of thermal stability at 90-40 Ma and from 20 Ma to present. The thermal history of Guangdong Province can be broadly divided into three phases. South of Gaoyao-Huilai Fault Zone, the thermal history exhibits one phase of relatively rapid cooling at 150-80 Ma, a period of thermal stability at 80-50 Ma, and two episodic cooling after 50 Ma. Between Gaoyao-Huilai Fault Zone and Xijiang River, the thermal history shows one phase of very rapid cooling at 120-80 Ma(reaching extremely rapid cooling in some areas), followed by a phase of very rapid cooling with minor temperature fluctuations at 30-10 Ma. North of the Xijiang,the timing of the first rapid cooling stage varies across different areas, occurring at 120 -100 Ma, 80-60 Ma, and 50-30 Ma. It is noted that there is a rapid cooling period around 10 Ma in all these areas. The thermal evolution of northern Guangdong-southern Hainan region since Late Mesozoic is likely to be related to Late Mesozoic subduction of Paleo-Pacific Plate, the Early Cenozoic rollback of Pacific Plate, and the Early Eocene collision of Indian-Eurasian Plates. Additionally, Hainan Island may have been further influenced by Philippine Plate and Paleo-South China Sea tectonics.

The Jurassic tectono-thermal evolution in northern Guangdong Province has generated abundant magmatic rocks, yet their petrogenetic mechanisms remain debated. This study investigates the magma source and its tectonic significance of Aizi granite in northern Guangdong by integrating LA-ICP-MS zircon U-Pb dating, trace element analysis, and zircon Lu-Hf isotopic data complemented by whole-rock Rb-Sr and Sm-Nd isotope analysis. The aim is to refine the genetic model for A-type granites formed under the Pacific Plate subduction setting during Late Jurassic. The Aizi granite yields a LA-ICP-MS zircon U-Pb age of 155±1 Ma(MSWD= 0. 12), indicating its formation during the Late Jurassic. The zircons exhibit negative εHf(t) values(-10. 4 to -4. 6), relatively high but slightly lower ( 87 Sr/ 86 Sr)i ratios(0. 713 1-0. 713 9) compared to the Mesoproterozoic basement of Cathaysia block, and zircon Hf crustal model ages(TDM2 = 1 490-1 856 Ma) that are older but marginally younger than Mesoproterozoic basement of Cathaysia block, suggesting that the magma source of Aizi granite is originated from ancient crustal materials with a small amount of mantle material added. The regional tectonic setting analysis indicates that Aizi granite is formed under the subsidence background of low-angle subduction of the Pacific plate slabs, triggered by asthenospheric upwelling that has induced partial melting of the upper crust in the weakened back-arc basin zones of Cathaysia ancient block, ultimately forming syenogranite.

Limestone slopes are highly susceptible to karstification and the detection of its karst characteristics is of great significance for elucidating water resource regulation mechanisms, karst water migration patterns, and ecological environment management in karst areas. A comprehensive method is conducted on the karst development characteristics of limestone slope in Chenqi, Puding County, Guizhou Province, integrating electrical resistivity tomography(ERT), ground penetrating radar(GPR), and geological drilling technology. A total of 7 resistivity inversion images at different positions of the slope are obtained through ERT, and the development degree of karst fissures is analyzed by integrating GPR data profiles, this research ultimately reveals the overall karst development characteristics of the study area. The results show that the karst development in limestone slope land exhibits significant spatial heterogeneity, with pipe-type karst predominantly developing in the middle of the slope land. Electrical resistivity tomography(ERT) effectively identifies the karst spatial distribution characteristics of limestone slope and by combining ground penetrating radar technology, it can further enhances the resolution of detecting karst development intensity and fracture development. Geological drilling technology is an effective mean to verify the applicability of these two methods. 

The application of geological big data in 3D geological structure modeling is explored, with Chengdu taken as an example. By utilizing the achievements of geological big data in Chengdu, this paper adopts a 3D modeling method of multi-source data fusion to construct 3D geological structure models at two scales of municipal administrative region and built-up area. Through 3D model data, it extracts four important geological structural planes and geological bodies such as faults, base-cover interfaces, fine-grained soil, and pebble soil. Their 3D spatial structure change characteristics are analyzed. The research findings can provide a scientific basis for land spatial planning, underground space development and utilization, and site selection for major projects in Chengdu.

The ecological texture serves as the carrier and identity of the distinctive features and cultural heritage of traditional villages around scenic areas. Its degradation is irreversible, making it particularly crucial to understand its vulnerability. Using vulnerability theory, this study establishes an analytical path and assessment framework for the ecological texture vulnerability of traditional villages around scenic areas. Based on vulnerability survey data and focusing on the three elements of exposure, sensitivity, and adaptability, the mechanism of ecological vulnerability is summarized and its connotation is clarified. By integrating field surveys, quantitative analysis methods, and spatial planning theories, and through analyzing the current survival environment and state of the villages, the spatial manifestations, causes, and external driving forces of their vulnerability are examined. From the two dimensions of external and internal influencing factors, the spatial logical association of ecological texture vulnerability in traditional villages around scenic areas is constructed, and a system of influencing factors for this vulnerability is deduced. The key factors for ecological texture protection lie in the superimposed external influence of the unique tourism-driven effects from the scenic area. Under this influence, multiple key factors are formed, including the ecological and social values of the village and tourist behavior. Compared to the relatively stable factors of physical spatial conservation, these factors more prominently highlight the ecological sensitivity of traditional villages around scenic areas, enhance the sustainability of conservation for such villages, and provide rational decision-making support through suitability assessment for their revitalization and utilization.

In order to solve the problem of UHPC shrinkage, magnesite calcination is used to prepare magnesia expansion agent(MEA), and the effects of calcination temperature-holding time and calcination-powder grinding sequential relationship on the activity of MEA samples are investigated. Based on the above results, MEA samples with different activities are prepared and mixed into ultra-high performance concrete(UHPC) for compressive strength and shrinkage performance testing. The results show that the decomposition rate of magnesite increased with the increase of calcination temperature and holding time. With the increase of calcination temperature and holding time, the crystallinity of MEA becomes higher, structure will be denser, and the activity decreases. Under the same calcination temperature and holding time, the structure of MEA prepared by the burnand-grind method is sparser and more active than that prepared by the grind-and-burn method. The compressive strength and shrinkage value of UHPC decreases with the increase in the activity of MEA. When the early autogenous shrinkage of UHPC doped with 6% of the highest activity, Ⅰ-MEA1 can be reduced by 33. 31%.

Aiming at the engineering problems of low strength and poor water stability of silt in the Yellow River area, a curing agent(SBCG) is prepared with steel slag, granulated blast furnace slag, and desulfurized gypsum as the main materials to improve and solidify the Yellow River alluvial silt. The road performance of SBCG-solidified silt is comprehensively evaluated by standard unconfined compressive strength, water immersion softening, freeze-thaw cycles, etc. , and the internal microstructure of solidified silt is observed by SEM. The results show that the incorporation of SBCG can significantly improve the road performance of silt. The solidified silt is easier to be compacted than plain soil, and the standard curing strength of each age is significantly improved with the increase of SBCG content. The loss rate of water immersion strength at 3 d age is less than 20%. The CBR after immersion still meets the subgrade strength requirements. The residual strength ratio after melting cycle is not less than 80%. According to the change law of strength of solidified silt under different test conditions, it is determined that the content of SBCG can be controlled below 10%. The SEM images show that SBCG can be hydrated inside the silt to form a large amount of hydrated calcium silicate gel(C -S -H), which is wrapped around soil particles, and ettringite(AFt) is filled in pores, making the silt form dense cementation. The structure is beneficial to the improvement of water stability and frost resistance.

To investigate the variation law of mechanical response of the graded crushed stone base stabilized with low cement content, the mechanical response characteristics of graded crushed stone structural layer stabilized with low cement content under different asphalt pavement structures are comparatively analyzed by using a meshless numerical simulation software. It is found that the mechanical parameters such as deflection and bottom stress of the new and reconstructed asphalt pavement structure changes with the difference in thickness and modulus of the low cement content stabilized graded crushed stone base. Thickness has a greater impact on bending and bottom stress than modulus, and the low cement mixing base is mainly compressive under different pavement structures. The results show that the low cement content of upper base layer improves the structural stress state of pavement, and the thickness should be controlled at about 20-30 cm, the modulus should be controlled at about 800 -1 000 MPa, and the cement content will be controlled at 2. 0%-2. 5%. In the case of ensuring the low cement content base under pressure, the deflection and tensile stress at semi-rigid structural layer should be used as the control index.

The strong thixotropy of Zhanjiang Formation structural clay causes the obvious time effect of pile foundation settlement, and there is a large error between measured value and calculated value obtained according to existing pile settlement timeliness calculation method, the applicability of existing pile foundation settlement timeliness calculation method is challenged in Zhanjiang Formation structural clay. Taking the pile foundation in Zhanjiang Formation structural clay as research object, pile foundation settlement time effect model experiment and pile side soil thixotropy experiment are carried out, the time effect coefficient of pile foundation settlement and thixotropic strength ratio of pile side soil at different standstill times are obtained, and the quantitative relationship between time effect coefficient of pile foundation settlement and thixotropic strength ratio of pile side soil is fitted. Based on hyperbolic model method, a calculation formula for time effect of pile foundation settlement considering thixotropy of Zhanjiang Formation structural clay is proposed, and the effectiveness of this calculation formula is verified by experiment and referenced data. The result shows this calculation formula can narrow the error between calculated and measured value of pile foundation settlement after a long-term load in actual project and provides a reference for the time effect calculation of pile foundation settlement in Zhanjiang Formation structural clay.

To achieve accurate monitoring of cable forces during the construction of bridge cables, a method is proposed where built-in polymer strain sensors are adopted to measure the cable forces of bridge cables in the construction stage during their fabrication process. Based on the basic principles of polymer strain method, the cooperative deformation of intelligent steel wires and cables are utilized to realize the measurement of cable forces. The packaging characteristics of polymer strain sensor are explained, and static load test cable force measurement results are compared with the traditional magnetic flux sensor measurement method. The results show that the polymer strain parallel wire intelligent cable produced by the polymer strain cable force measurement technology has good sensing linearity and repeatability, and the identifiable cable force error of intelligent cable is -1. 38%, which has high accuracy and reliability. The measurement accuracy meets the requirements, and is suitable for rapid cable force monitoring in construction stage.

Taking the nonuniform surface subsidence data of the Panyu Plaza construction section of Guangzhou subway line 22 as an example, autoregressive moving average(ARIMA) model and Gaussian process regression (GPR) model are used to analyze and predict the change of surface subsidence. Firstly, the training set data is used to establish ARIMA model, determine the ARIMA(3, 1, 2) model and predict its linear subsidence, and then GPR model is used to train the ARIMA model residual and predict its nonlinear value. The predicted value of surface subsidence is obtained by adding the predicted values of the two models. The results show that the prediction error of ARIMA-GPR mixed model is significantly reduced compared with ARIMA and GPR single prediction models, and the prediction accuracy is improved by above 40%, which improves the prediction accuracy of single prediction model. Therefore, ARIMA-GPR mixed model is an effective and feasible method for nonuniform subsidence prediction.

With the gradual increase in the scale of substation construction and the types of transmission and transformation components, it is difficult to meet the needs of massive 3D data rendering of substations only relying on the upgrading of computing and storage capabilities. Optimizing the scheduling of 3D scene data is an indispensable means to improve rendering efficiency. Through analyzing the spatial storage structure of loose octree and combining with visual correlation dynamic LOD technology, a dynamic scheduling method of substation scene model based on loose octree is proposed. This method takes the sight distance as the constraint condition, indexes the loose octree nodes from top to bottom, simplifies the objects in the qualified nodes to the corresponding LOD level, and improves the rendering rate of 3D scene of the substation by reducing the amount of model data. Finally, taking a substation project as the research object, a fast rendering of 3D substation scene is realized based on Unity 3D, which verifies the effectiveness of the dynamic scheduling method.

3D electrolysis is applied to treat biochemical tail water from wastewater treatment plant, and the chemical oxygen demand(COD) removal rate in wastewater is investigated. The effects of operating parameters (voltage, initial pH, distance between electrodes, and electrolysis time) on COD removal rate are analyzed with response surface methodology(RSM). Optimization of operating parameters is carried out with the central combination Box Behnken Design(BBD) and a mathematical model for COD removal rate is established. The results show that the order of effect of interaction between various factors on COD removal rate is distance between electrodes> initial pH> voltage > electrolysis time. Under the optimized operating conditions( voltage = 20. 13 V, initial pH = 4. 85, distance between electrodes = 7. 65 cm, electrolysis time = 125. 16 min), the best COD removal rate is 55. 94%. In the validation experiment, COD removal rate is 53. 40%, with only 2. 54 percentage points deviation from the predicted value, indicating the effectiveness of RSM optimization.

The calcite-otavite solid solution(Ca1-xCdx)CO3 is synthesized and its dissolution in air-saturated water at 25 ℃ for 270 d is experimentally explored. During dissolution, the aqueous Cd 2+ concentration increases to the peak value of 0. 082-11. 102 μmol / L in 6 h, and then declines slowly to stable state of 0. 026-1. 334 μmol / L after 210-270 d dissolution. Cd 2+ / (Ca 2+ +Cd 2+ ) mole ratio of the aqueous solution is greatly lower than Cd / (Ca+Cd) mole ratio(XCd ) of the corresponding solid showing a strong non-stoichiometric release of calcium and cadmium. At 25 ℃ , the solubility product(Ksp ) and Gibbs free energy of formation(ΔG ° f ) are determined to be 10 -8. 40±0. 02 to 10 -8. 40±0. 01 and - 1 129. 40 ± 0. 11 to - 1 129. 38 ± 0. 04 kJ/ mol for calcite ( CaCO3 ), 10 -11. 59±0. 07 to 10 -11. 51±0. 17 and -671. 61 ± 0. 45 to - 671. 18 ± 0. 99 kJ/ mol for otavite (CdCO3 ), respectively. With the increasing XCd , the Ksp value declines non-linearly from 10 -8. 40±0. 02 to 10 -8. 40±0. 01 for calcite to 10 -12. 13±0. 09 for (Ca1-xCdx)CO3 solid of XCd = 0. 90, and then increases slightly to 10 -11. 59±0. 07 to 10 -11. 51±0. 17 for otavite; while the ΔG ° f value increases linearly. Lippmann diagram is constructed for (Ca1-xCdx)CO3 solid solution with the estimated Guggenheim coefficients a0 = -2. 36 and a1 = -6. 89. The solid dissolves non-stoichiometrically and the aqueous solution moves progressively up to solutus curve and the minimum stoichiometric saturation curve for pure otavite, from right to left, finally approaches the intersection region of stoichiometric saturation curves for pure calcite and (Ca1-xCdx)CO3 solid solutions(XCd = 0. 21-0. 90). The substantially Cd-poor aqueous solution is lastly in equilibrium with CdCO3 -rich solid.

From the perspective of ecological and social coupling, taking Guilin as an example, based on ecological security human footprint, this study constructs an ecological security pattern through methods such as the minimum cumulative resistance model and circuit theory, and identifies key areas for ecological structure and functional damage in territorial space restoration. (1) Human activity interference is concentrated in Guilin, with high interference areas accounting for 8. 95% of the total area of study area, and is distributed in the built-up areas of Diecai, Xiufeng, Qixing, Xiangshan, eastern Lingui, and other counties. (2) The overall ecological security pattern in Guilin presents a network pattern feature, which includes 13 ecological source areas and 25 potential ecological corridors among sources. (3) Based on human interference assessment and ecological security pattern, the key areas of 1 787. 76 km 2 for ecological source restoration, 124. 49 km for ecological corridors restoration, and 74 key ecological pinch points restoration areas are identified, mainly distribute in Yangshuo, Ziyuan, Xing??an, and other areas. (4) From the perspective of combining area-line-point, it proposes targeted restoration strategies for three types of key ecological resoration areas, focusing on integrating fragmented source patches, creating green corridors, and improving road environment. This research can provide theoretical basis and decision-making reference for current ecological restoration work of territorial space.

In order to meet the current requirements for paint surface defects detection in intelligent factories and solve the problems of uneven light intensity, small defects, and large disparity in number of positive and negative samples during paint surface defects detection, a Copy-Paste method based on pixel clustering constraints data augmentation method and MiniYOLO lightweight algorithm are proposed in this paper. Firstly, this paper performs pixel clustering on each picture of paint surface data set under light intensity, and enriches the defect characteristics of data set by copying and pasting target defects under the same light conditions. Then a micro-scale detection layer is added based on original YOLOv5s. For network refinement, K-means++ algorithm is used to regenerate priori box to improve model detection ability of small targets. Finally, coordinate attention mechanism is used in backbone network to obtain MiniYOLO to eliminate false detection of non-defects in paint surface that resember defects. Experimental results show that mAP of MiniYOLO in this paper reaches 97. 7% on paint surface data set, and model size is only 40. 3 MB. Therefore, it can effectively meet paint surface inspection needs of actual industrial production.

Aiming at the problems of smooth edges and fuzzy details in existing real image denoising, an end-toend denoising algorithm based on spatial and channel self-attention is proposed. In feature extraction layer, spatial and channel self-attention are combined to enhance the network’s ability to understand different regions and diverse features. The feedforward neural network with improved self-attention is adopted to enhance the network’s capability of understanding local image context. Residual learning is employed to apply the improved feature extraction layer into encoder-decoder architecture with multi-scale feature fusion. The experimental results show that the peak signal-to-noise ratio of SIDD and DND data sets are 40. 52 and 40. 12 dB, and the structural similarity is 0. 972 and 0. 967 respectively, which effectively improves the denoising effect and preserves more detail texture and edge structure information of images.

Taking 31 districts and counties of 5 prefecture-level cities along Guangxi Border and Coastal National Scenic Route as study area, GIS spatial analysis, tourism development index analysis, accessibility analysis, and gravity model analysis are used to identify tourism spatial structure of National Scenic Route. (1)Tourism resources exhibit an enriched-central-east but sparse-northwest spatial layout, while tourist facilities show a higher concentration in central region compared to the west; both demonstrate clustered spatial distribution patterns. (2)Daxin County, Dongxing City, and Haicheng District are identified as primary tourism nodes. Hepu County, Qinnan District, Qinbei District, Fangcheng District, Port District, Ningming County, Pingxiang City, Jiangzhou District, Longzhou County, Jingxi City, and Tianyang District are identified as secondary tourism nodes. (3)Based on buffer analysis, tourism resource accessibility analysis, and transportation network accessibility analysis, seven tourism development corridors are identified(one primary, four secondary, and two tertiary). (4)The tourism economic linkages along the route display a pattern of denser in the east and stronger in the south, sparser in the west and weaker in the north, overall forming a network-belt spatial structure. Recommendations for optimizing the tourism spatial structure of Guangxi Border and Coastal National Scenic Route are proposed from the perspectives of tourism nodes, corridor design, and economic linkage.

Tourism development can promote the integration of all ethnic groups in social, economic, cultural, psychological, spatial and other aspects, and plays an important role in strengthening the construction of Chinese national community and forging a strong sense of the Chinese national community. Based on the theory of cognitive - affection - intention, this study builds a structural equation model of tourism impact perception, tourism support attitude, interaction, exchange and integration attitude and sense of community for the Chinese nation based on the research area of Kashgar Ancient City scenic spot. The results show that: (1) tourism positive impact perception, tourism support attitude, interaction, exchange and integration attitude have significantly positive effects on forging a strong sense of community for the Chinese nation. (2) Tourism support attitude and interaction, exchange and integration attitude play a chain intermediary role in the process of tourism impact perception on forging a strong sense of community for the Chinese nation. (3) Tourism development cost perception has no significant impact on tourism support attitude, interaction, exchange and integration attitude and sense of community for the Chinese nation