Table of Content

25 November 2024, Volume 0 Issue 11

Previous Issue    Next Issue

Inversion of soil moisture in the Yuanmou hot-dry river valley area based on the PSO_GA-RBF neural network model

DU Jinming, LUO Mingliang, BAI Leichao, WU Qiusheng

2024, 0(11):  1-6.  doi:10.13474/j.cnki.11-2246.2024.1101

Asbtract ( )   PDF (2083KB) ( )  

References | Related Articles | Metrics

Soil moisture has a significant impact on hydrological and climatic processes. A comprehensive and accurate understanding of soil moisture status is of great research value for hydrological simulation, ecological governance, and other related fields. In response to the soil moisture inversion issue in the Yuanmou hot-dry river valley area, a new soil moisture inversion model is constructed using the PSO_GA-combined optimized RBF neural network. The experiment utilizes Sentinel-1 radar data and Sentinel-2 optical data, and employs the water-cloud model suitable for low vegetation cover types in the study area to correct the vegetation scattering effects. The obtained VV and VH polarized soil backscattering coefficients and cross-polarization differences are incorporated into the constructed model, enabling the remote sensing inversion of soil volumetric water content in the hot-dry river valley area of Yuanmou county, Yunnan province. Comparisons and validation against measured soil volumetric water content data show a root mean square error of 0.55% m³/m³ and a coefficient of determination (R²) of 0.855, demonstrating a significant improvement in accuracy compared to traditional RBF neural network models.Correlational analysis is conducted between the inversion results and NDVI values, revealing a coefficient of determination (R²) of 0.512 7 between the two. This verifies the high precision of soil volumetric water content inversion based on Sentinel-1 radar image data, utilizing the water-cloud model and PSO_GA-combined optimized RBF neural network, validating the feasibility of large-scale soil moisture monitoring in hot-dry river valley areas.

Segmented bundle adjustment algorithm for underwater vision SLAM

BAI Yunpeng, XU Huixi, Lü Fengtian

2024, 0(11):  7-12.  doi:10.13474/j.cnki.11-2246.2024.1102

Asbtract ( )   HTML ( )   PDF (1722KB) ( )  

References | Related Articles | Metrics

Autonomous underwater vehicle (AUV) can achieve close-range accurate positioning by using visual SLAM system,but when facing large-scale underwater scenes,the back-end optimization using bundle adjustment(BA) algorithm has the problems of insufficient memory and low computational efficiency. To solve these problems,an improved segmented BA optimization algorithm is proposed. A segmentation method based on motion pattern is used to segment the trajectory according to the straight motion and turning motion of the camera,and then BA optimization is performed on each sub-segment respectively. Each sub-segment is solved by dynamically adjusting the optimization weight,and the optimization parameters are dynamically adjusted according to the motion patterns of different sub-segments. For the solving of BA cost function,the improved Levenberg-Marquadt(L-M) algorithm is adopted,the trust region is defined as the tunable parameter,which reduces the non-convergence problem caused by the singularity of the Jacobian matrix and improves the operation efficiency. According to the experimental results on the dataset,the proposed algorithm has better accuracy than the ORB-SLAM3 algorithm when it runs for a long time and the environment is harsh,and the efficiency of the global BA is significantly improved.

Spatio-temporal inversion of soil salinity in the Yellow River Delta region based on GEE

FU Pingjie, BU Yuankun, MA Chijie, LI Xiaotong, MA Mingliang

2024, 0(11):  13-20,26.  doi:10.13474/j.cnki.11-2246.2024.1103

Asbtract ( )   HTML ( )   PDF (3066KB) ( )  

References | Related Articles | Metrics

In this study, the Yellow River Delta is taken as the research area. Based on the remote sensing image and ground measured salt data in 2022, the remote sensing spectral index and band reflectance with strong correlation with salt data are extracted as modeling factors. Multiple linear regression, random forest, BP neural network and XGBoost regression method are used to construct soil salt inversion model, and the optimal model is selected to carry out long-term inversion analysis of soil salt content in the study area from 2001 to 2020. The results show that: ①Through correlation analysis, 8 spectral information(CRSI, DVI, ENDVI, MSAVI, NDSI, NDVI, SI-T, near infrared band)related to soil salt content are screened out, which are significantly correlated at the level of P<0.01. ②Compared with the prediction accuracy of the four inversion models, the XGBoost algorithm has a stable prediction ability, and the inversion effect of soil salinity in the study area is the best. The values of R² and RMSE in the validation set are 0.84 and 3.066. ③According to the soil salt content from low to high, the salinization grade of the study area is divided into four grades (Ⅱ, Ⅲ, Ⅳ and Ⅴ). In the past 20 years, the total area of saline soil in the study area showed a downward trend, reducing by 20.7% of the total area of the study area.

Automatic extraction of water body using multi-feature index

ZHANG Baowen, ZHAO Zhan

2024, 0(11):  21-26.  doi:10.13474/j.cnki.11-2246.2024.1104

Asbtract ( )   PDF (3268KB) ( )  

References | Related Articles | Metrics

At present, water extraction is still a semi-automatic method, which has low efficiency and is prone to omissions and extraction errors. In this paper, an automatic water extraction method using multi-feature index is proposed. Vegetation index, water index, improved water index and building land index are used to automatically extract water and non-water samples, and the classifier is trained to realize high-precision and automatic urban water extraction. Four typical experimental areas with different geographical conditions at home and abroad are selected for comparison and analysis with the traditional extraction method. Under this method, high extraction accuracy is achieved in all the four experimental areas, and Kappa coefficient reached above 0.94. Compared with single-band threshold method, Kappa coefficient increases by 7.2% on average. Compared with the water index method, Kappa coefficient increased by 3.8% on average.

Automatic extraction of transmission line flat sections and vegetation envelopes based on airborne laser point clouds

ZU Weiguo, TAN Jinshi

2024, 0(11):  27-32.  doi:10.13474/j.cnki.11-2246.2024.1105

Asbtract ( )   PDF (2077KB) ( )  

References | Related Articles | Metrics

The flat cross-section map of transmission lines is a crucial basis for optimizing transmission line routes, reducing environmental impact, and lowering costs. Traditional methods for mapping flat cross-sections involve considerable labor and high risk, with low accuracy in estimating vegetation height and rare direct vegetation envelope lines to affect line design. To address this, a method for automatically extracting and drawing flat cross-sections and vegetation envelope lines of transmission lines based on airborne laser point clouds is proposed. The overall technical approach are described firstly. Then, key technologies are investigated in depth, including laser point cloud calculation and DSM construction, filtering and DEM construction, automatic extraction of flat cross-sections with adaptive intervals, and vegetation envelope line simplification algorithms. Finally, through practical cases, the DSM and DEM construction from laser point clouds, extraction of flat cross-sections and vegetation envelope lines, and their accuracy and efficiency are analyzed. Results show that laser point clouds provide high measurement accuracy, enabling fast, accurate, and automated extraction of flat cross-sections and automatic drawing of vegetation envelope lines, So offer significant technical support for optimizing transmission line routes and being worthy of widespread promotion and application.

Attitude monitoring method of mooring ship based on multi-line LiDAR

YU Wei, CAO Min, SUN Jinyu, WU Jiwei, TIAN Jin, HUANG Xiusong

2024, 0(11):  33-37.  doi:10.13474/j.cnki.11-2246.2024.1106

Asbtract ( )   HTML ( )   PDF (1637KB) ( )  

References | Related Articles | Metrics

To achieve real-time monitoring of container ship mooring and loading/unloading operations, ensuring operational safety and enhancing efficiency, we introduce the near-neighbor point cloud center of gravity sampling algorithm, the warp/latitude scanning boundary extracting algorithm based on the distance threshold, and the attitude angle algorithm of container ship. Additionally, we construct a simulation scene model and establish a dataset of motion point clouds for moored container ships to validate the effectiveness of these algorithms. Comparison between calculated and actual values from simulated data collected by radars with varying line counts demonstrates the algorithm's accurate reconstruction of the ship's motion history. The measurement error for transverse/vertical rocking inclination is less than 0.2°, and the algorithm's sensitivity to point cloud density within the effective range is negligible. This initial validation confirms the algorithm's effectiveness and feasibility, offering data support for real-time monitoring and early warning systems for container ship buoyancy and stability.

Airborne realistic 3D reconstruction based on visual pose correction

GAO Yunhan, LIN Yili, ZHANG Jinghan, QIAN Wei, XING Yubo, SHI Hang, XIE Yangmin

2024, 0(11):  38-43.  doi:10.13474/j.cnki.11-2246.2024.1107

Asbtract ( )   HTML ( )   PDF (1845KB) ( )  

References | Related Articles | Metrics

In urban surveying and mapping, drones equipped with LiDAR hover to collect data are prone to motion distortion due to the influence of fuselage vibration, resulting in poor fusion modeling performance. This article proposes a three-dimensional reconstruction method for airborne laser real scenes based on visual pose correction. The method utilizes video information obtained from binocular cameras during the scanning process of LiDAR to process the trajectory of LiDAR pose changes to correct the pose of the laser point cloud. Then, the corrected laser point cloud is projected onto the coordinate system of a monocular camera, and color information is obtained through collinear equations to achieve information fusion and generate a three-dimensional true color point cloud, and evaluation criteria for four feature dimensions of straightness, flatness, verticality, and fusion coloring are established for the true color point cloud before and after visual pose correction. The experimental results show that the true color point cloud after visual pose correction has a maximum improvement of 77.3% in straightness, 54.5% in flatness, and 54.53° in verticality compared to before correction. The color attachment is significantly more accurate.

Monitoring and prediction of ground subsidence in mining areas using DS-InSAR and LSTM

WANG Benhao, WANG Yanxia, XIANG Xueyong, HU Hong

2024, 0(11):  44-48.  doi:10.13474/j.cnki.11-2246.2024.1108

Asbtract ( )   PDF (3558KB) ( )  

References | Related Articles | Metrics

In response to the problem of low point density and uneven distribution in subsidence monitoring of mining areas using conventional InSAR technology, this paper uses 36 Sentinel-1A image data from August 2020 to August 2023 to obtain surface deformation information of Langyashan mining area in Chuzhou city, Anhui province using DS-InSAR technology. And the LSTM neural network model is used to predict the future settlement trend of the area with severe ground subsidence in the mining area, in order to understand the future development trend of ground subsidence in the mining area. The research results indicate that:①Compared with traditional PS-InSAR technology, DS-InSAR technology can significantly increase the number of monitoring points in mining areas and more comprehensively reflect surface subsidence information in mining areas. ②During the monitoring period, there are three deformation zones in the mining area, with a maximum settlement of 32.4 mm and a maximum settlement rate of 10.8 mm/a. ③By comparing with the GM (1,1) model and using the selected 6 settlement feature points, it is found that the LSTM neural network model exhibited higher prediction accuracy. ④For the area with the highest cumulative settlement, we use the LSTM model to predict the cumulative settlement of the 6 feature points in the area for the next 12 months. The prediction results show that the future settlement in the area fluctuates within a certain range, and no obvious settlement trend has been observed yet.

Correction of vehicle-borne laser point clouds in complex urban environments based on POS

HAN Xuan, LIU Rufei, CUI Jianhui, WANG Minye, LI Zeyu

2024, 0(11):  49-55.  doi:10.13474/j.cnki.11-2246.2024.1109

Asbtract ( )   HTML ( )   PDF (4572KB) ( )  

References | Related Articles | Metrics

In the urban environment with tall buildings and dense trees, the GNSS signals are susceptible to occlusion, and the position error of the POS composed of GNSS/INS accumulates rapidly. This leads to an increase in the coordinate errors of vehicle-borne laser point clouds computed from the position attitude data provided by the POS. Additionally, the point cloud shows localized non-rigid deformation phenomena. To solve the above problems, based on the traditional POS-based point cloud correction method, this paper further studies the complex urban environment with GNSS signal occlusion and vehicle operation conditions. Through comparing the spacing of different control points, it formulates the optimal deployment scheme for control points. The error characteristics of POS under different GNSS signal occlusion conditions are deeply analyzed so as to construct a reasonable error time-varying interpolation model. The control point error interpolation information is utilized to correct the POS position to provide reliable position information for the vehicle-borne laser point cloud solving. The experimental results show that the positioning accuracy of the POS corrected using the strategy of this paper can be improved by about 62.50% compared with the pre-correction one. The accuracy of the corrected point cloud solved using the corrected POS position can be improved by about 75.45% compared with the pre-point cloud correction one.

Implementing the pan-map dimension association approach for trajectory visualization

DENG Zhigang, GUO Renzhong, CHEN Yebin, MA Ding, ZHAO Zhigang, ZHU Wei

2024, 0(11):  56-60,96.  doi:10.13474/j.cnki.11-2246.2024.1110

Asbtract ( )   PDF (1419KB) ( )  

References | Related Articles | Metrics

The barrier to participation in trajectory expression has continuously lowered with the development of information technology; however, the trajectory expression scheme still heavily relies on the user's comprehension of the information value and expression requirements. As a result, this study establishes the association between needs and visualization dimensions based on the theory of visualization dimensions and presents the generalization expression requirements of trajectory data in light of the generalization characteristics of trajectory expression. The space-time cube, personal travel timeline, and OD trajectory flow diagram are used as examples to analyze the trajectory expression characteristics under various dimensional combinations. The research focuses on the analysis of the expression characteristics between the temporal structure, geometric logic, spatial dimension, spatial organization, and expression scale dimensions and the expression demand. The findings demonstrate that it is possible to do trajectory expression using the visual dimension theory, and that studying the visualization dimension theory's trajectory expression is very beneficial.

Improved object detection algorithm HCAM-YOLO in traffic scenes based on YOLOv5

WANG Zhitao, ZHANG Ruiju, WANG Jian, ZHAO Jiaxing, LIU Yantao

2024, 0(11):  61-67.  doi:10.13474/j.cnki.11-2246.2024.1111

Asbtract ( )   PDF (2352KB) ( )  

References | Related Articles | Metrics

The rapid and precise detection of targets in traffic scenarios is crucial for intelligent traffic management and driving path decision-making. Traditional target detection models often grapple with issues such as inadequate detection accuracy, high rates of leak detection and false detection due to the complexity and variability of the traffic environment, and the diversity and sparsity of target features. To address these challenges, this paper introduces a YOLOv5 target detection model, HCAM-YOLO, which leverages the HcPAN feature fusion network. The crux of this approach lies in addressing the issue of local information being easily lost during the PAN network's feature fusion process. A hybrid convolutional attention mechanism(HCAM) is designed to enhance multi-scale information extraction in feature fusion networks. By integrating the HCAM module into the PAN's underlying structure, the sensitivity of key local features is enhanced, while the fusion effect of deep semantic information and shallow positional data is strengthened. This method's novelty lies in its use of an attention mechanism to optimize the feature fusion process, thereby improving the model's detection performance of pedestrians, motor vehicles, and other targets in complex traffic environments. The experimental dataset comprises the Rope 3D dataset, Road Veh dataset, and Road Ped dataset. The results demonstrate that the HCAM module is more suitable for integration into the underlying PAN network than other attention mechanisms. When compared to the basic YOLOv5 model, the precision and recall of the final HCAM-YOLO algorithm model increased by 3.4% and 3.2%,respectively, and mAP@0.5/% by 3.8%. The HCAM-YOLO algorithm model proposed in this paper exhibits strong adaptability to target detection tasks in traffic scenes with complex backgrounds.

Color restoration of underwater images using color compensation and convolutional neural network based defogging model

MA Zhenling, CHEN Yuan, FAN Chengcheng, PAN Yan

2024, 0(11):  68-73.  doi:10.13474/j.cnki.11-2246.2024.1112

Asbtract ( )   PDF (1834KB) ( )  

References | Related Articles | Metrics

Underwater vision measurement has important applications in marine surveying, underwater engineering surveying, underwater archaeology and underwater environmental monitoring. However, underwater images suffer from color distortion, image blurring and low contrast, which limits the application of underwater visual measurement technology in practical environments. A color restoration method for underwater images based on color compensation and convolutional neural network (CNN) defogging model is proposed in this paper, in which the image enhancement is carried out step-by-step.Firstly,the color deviation of underwater images is analyzed, and then an adaptive color compensation strategy combined with the grayscale world white balance algorithm is used to correct underwater image color. Secondly, a CNN based dehazing model was designed to achieve dehazing processing of underwater images. Finally, the adaptive histogram equalization CLAHE method is used to enhance the contrast of underwater images. In order to prove the applicability and superiority of the proposed method, two image datasets are combined to study, and several known underwater image enhancement and restoration methods are compared. The proposed method and several compared methods are evaluated in two aspects of subjective visual effect and quantitative evaluation index. The comparison results show that compared with other enhancement algorithms, the proposed method successfully improves the clarity of the image and reduces the color deviation of the damaged underwater image when processing underwater images in various environments and has superior image color recovery compared with existing enhancement methods.

Application of landslide geohazard investigation based on realistic 3D and GIM technology

MA Jianxiong, MING Jing, ZHOU Chengtao, GAN Ze

2024, 0(11):  74-77,161.  doi:10.13474/j.cnki.11-2246.2024.1113

Asbtract ( )   PDF (3612KB) ( )  

References | Related Articles | Metrics

Accompanied by the accelerating pace of infrastructure construction and the continuous improvement of the quality requirements of engineering construction, the accuracy of landslide engineering survey also puts forward higher requirements, this paper takes a project on the right bank of the Yangtze River in Liangjiang New Area of Chongqing Municipality as an example, and adopts the geological fine investigation method based on the realistic 3D and GIM technology, realizing a breakthrough of the application technology of three-dimensional fine survey of landslide engineering in the mountainous city, which provides support for the emergency disposal of landslide and the implementation of engineering. The results show that:① Based on the real-life 3D model realized the interpretation of the key elements of landslide geohazards, which provides a powerful auxiliary decision-making and basis for geohazard investigation; ② Through the 3D geologic modeling method with GIM technology as the core, the construction of a refined 3D geologic information model of the study area is realized, which has the characteristics of rapidity, accuracy, and reliability; ③ Through the integrated integration and fusion application of the real-life 3D model and GIM technology, the survey results are more refined, providing effective data support and scientific decision-making for the emergency disposal of landslides and the implementation of he project.

GNSS/INS integrated navigation algorithm with UWB constraints

ZHOU Tao, ZOU Jingui, ZHAO Yinzhi, ZHOU Zhennan, WU Jingwen, HUANG Junfeng, MIN Hui

2024, 0(11):  78-82.  doi:10.13474/j.cnki.11-2246.2024.1114

Asbtract ( )   PDF (2258KB) ( )  

References | Related Articles | Metrics

In order to solve the problem of traditional integrated navigation algorithms being difficult to apply in satellite signal rejection and indoor environments,ultra wideband (UWB) positioning technology is considered to assist GNSS/INS navigation systems. We derive a combined navigation Kalman filter model with additional UWB constraints and construct a noise matrix using the prior variance of UWB distance measurement at the data processing level. To verify the effectiveness of the proposed method, we conduct a trolley experiment in the air raid shelter of Wuhan university. The experimental results show that the addition of UWB positioning information can effectively suppress the rapid divergence of velocity and position. After 50 seconds of GNSS signal interruption, the velocity drift in the E, N, and U directions is 0.236, 0.284, and 0.179 m/s,respectively, and the position drift is 5.247 m. The navigation positioning accuracy is improved by more than 80% compared to traditional algorithms.

A study on relative positioning of BDS considering GEO orbital characteristics

WANG Tianle, GAO Chengfa, WEI Jianxiong

2024, 0(11):  83-89.  doi:10.13474/j.cnki.11-2246.2024.1115

Asbtract ( )   PDF (3039KB) ( )  

References | Related Articles | Metrics

Considering the issue that dynamic relative positioning of the BDS system rarely takes GEO orbital characteristics into account, this paper analyzes the spatial signal quality and observation value quality of GEO satellites. By considering GEO orbital characteristics, the stochastic model is improved based on inter-station single-difference residuals and a downweighting factor. The positioning performance is verified and analyzed under the conditions of excluding all GEO satellites and excluding some GEO satellites. The experimental results show that considering the GEO orbital characteristics, excluding some GEO satellites, and improving the stochastic model can ensure the number of observed satellites, enhance the ambiguity resolution rate, and improve the accuracy of dynamic relative positioning.

Application of topographic survey of island (reef) based on multi-beam sounding system

SUN Dong, DING Shijun, LI Xiaohong, LIU Yuan, LIU Haibin

2024, 0(11):  90-96.  doi:10.13474/j.cnki.11-2246.2024.1116

Asbtract ( )   PDF (2142KB) ( )  

References | Related Articles | Metrics

With the deepening of marine comprehensive survey, islands (reefs) as an important element of the ocean, its complete surface and underwater terrain data is the basis of understanding and planning islands and reefs. The multi-beam sounding system combined with 3D laser is used to obtain the integrated topographic data of islands and reefs above and below water. The experiment is carried out in combination with a typical island in eastern Shandong province, focusing on the application of the multi-beam sounding system in the topographic data acquisition of islands and reefs, the accuracy assessment is carried out, and the data results are displayed. The results of the study area show that the ship-borne measurement system combined with the unmanned ship measurement system can obtain the land and water interface area of the reef completely at one time by using the high-low tidal range, and the data is complete and high precision. The full coverage data results can truly reflect the topography of islands and reefs and their surrounding areas, which provide the data support for planning and construction and ecological monitoring.

Research and application of Lutan-1 SAR satellite in survey and monitoring of catastrophic geohazards

YU Zhonghai, YAN Libo, LIU Qian, LU Guangbo, LIU Rui

2024, 0(11):  97-101,176.  doi:10.13474/j.cnki.11-2246.2024.1117

Asbtract ( )   PDF (2931KB) ( )  

References | Related Articles | Metrics

Lutan-1(LT-1) is the first L-band differential interferometric SAR satellite in China. Jinan has established a satellite-based monitoring network to deepen the construction of comprehensive monitoring and early warning system for urban safety risks since 2024. The SAR satellites are designed for monthly deformation monitoring of major infrastructures such as bridges, super high-rise buildings, mines, and geological hazards. This paper conducted a surface deformation study based on LT-1 with LandSAR software for a 2800 km² area in the southern region the LT-1 is effective for geological hazard deformation survey and monitoring. At the same time, of Jinan. Research results showed that the obvious subsidence in some mining areas is also detected, which could provide monitoring basis for the supervision of production safety in mining areas.

Analysis of spatiotemporal evolution characteristics of typical landslides in the Jinsha River Basin based on SBAS-InSAR technology

YANG Fang, DING Renjun, LI Yongfa

2024, 0(11):  102-107.  doi:10.13474/j.cnki.11-2246.2024.1118

Asbtract ( )   PDF (2131KB) ( )  

References | Related Articles | Metrics

The Jinsha River Basin belongs to the high mountain canyon area, with complex geological and geomorphological features.The numerous canyons, steep terrain, and a large amount of rainfall have led to frequent landslide disasters,which have caused serious impacts on human safety,production,and the environment.However,conventional measurement methods have drawbacks such as high cost,long cycle,and insufficient spatial resolution,making it difficult to fully reflect the evolution characteristics of landslides.Therefore,this article uses SBAS-InSAR technology combined with Sentinel-1A data from the lifting track to obtain surface deformation information of the Ahai Reservoir area in the Jinsha River Basin from January 2019 to December 2020.Three typical landslides,namely Ligu,Baiya,and Luoziru,are selected for spatiotemporal evolution characteristics analysis.The research results indicate that SBAS-InSAR technology can effectively identify typical landslides in high mountain canyon areas.During the monitoring period,the maximum deformation rate of the Li Gu landslide is -68 mm/a,and the cumulative deformation variable is -148 mm.The overall spread from the deformation center to the west towards the Jinsha River is in a strip shape.The maximum deformation rate of Baiya landslide is -40 mm/a,and the cumulative deformation variable is -77 mm.The maximum deformation rate of Luoziru landslide is -90 mm/a,and the cumulative deformation reaches -260 mm.

Target detection method of railway catenary components in UAV images based on improved YOLOv7

SONG Zongying, WANG Xingzhong, ZENG Shan, ZHANG Zhengjun, YIN Taijun, LIU Hongli

2024, 0(11):  108-114.  doi:10.13474/j.cnki.11-2246.2024.1119

Asbtract ( )   PDF (2863KB) ( )  

References | Related Articles | Metrics

The catenary system, as an essential component of electrified railways, provides energy to trains and ensures their normal operation. Damage to catenary system components poses a threat to train safety, making it crucial to regularly inspect the condition of these components. In recent years, UAVs have been widely used in monitoring the condition of critical catenary components. However, due to the complex and variable backgrounds, significant scale changes, and the presence of many small targets in the catenary images captured by UAVs, existing detection algorithms frequently suffer from false detections and missed detections of catenary components. To address this issue, this paper proposes a catenary component target detection method based on an improved YOLOv7 algorithm. By introducing an enhanced receptive field module, the network's feature extraction capability is strengthened, leading to more discriminative target feature representations. Additionally, an improved coordinate attention mechanism is incorporated during the fusion of adjacent scale feature maps to highlight the target features of catenary components and suppress redundant background information. The bounding box loss function is optimized using the Wasserstein distance, effectively improving detection accuracy. Experiments on the catenary component dataset show that the improved YOLOv7 algorithm can accurately detect various catenary components in drone-captured images, achieving a mean average precision of 97.27%, which is 3.83% higher than before the improvement. The proposed algorithm enhances the high-precision and rapid detection capabilities of drones for critical catenary components, providing technical support for achieving more intelligent drone inspections.

Classification and modeling of 3D point clouds in offshore waters based on color and image guidance

YU Rui, ZHANG Huiran, LIAO Jianbo

2024, 0(11):  115-119.  doi:10.13474/j.cnki.11-2246.2024.1120

Asbtract ( )   PDF (1829KB) ( )  

References | Related Articles | Metrics

In order to solve the problems of accuracy loss and blurring the boundaries between sea surface and mudflat in coastal sea features classification, we proposes a 3D point cloud classification and modeling method in offshore waters through color and image-guidance strategies. This method first uses the color distribution histogram to guide the rough classification of ground objects, and then uses SVM to achieve fine classification of ground objects. The 2D classification results will then be optimally mapped to the 3D point cloud, and on this basis, 3D model reconstruction of islands, mudflats and other features is carried out. The experimental result indicated the proposed method can accurately and reliably classify offshore sea features and achieve 3D model reconstruction of surface features.

Study on multisource data fusion methods and their application in comprehensive subsidence monitoring of mining area surface

DU Yuzhu, LIANG Tao

2024, 0(11):  120-125.  doi:10.13474/j.cnki.11-2246.2024.1121

Asbtract ( )   PDF (2039KB) ( )  

References | Related Articles | Metrics

With the development of unmanned aerial vehicle (UAV), sensor, and data processing technologies, lightweight and low-cost UAVs can carry a variety of sensors to obtain diverse high-precision observation data. In response to the characteristics of mining-induced subsidence, this paper designs a lightweight and small-scale UAV mining area ground monitoring scheme that integrates aerial photography and LiDAR. It studies key technologies such as multi-period and multi-source data registration, selection of subsidence monitoring points, construction of surface rock movement observation lines, and proposes effective solutions. According to the research results, application tests have been carried out, and the results show that the lightweight UAV measurements using fused point clouds and imagery can obtain comprehensive mining area ground subsidence models with a precision better than 0.25 meters.

Monitoring of deformation in mining buildings based on GNSS and InSAR technologies

ZHAO Qi

2024, 0(11):  126-132,166.  doi:10.13474/j.cnki.11-2246.2024.1122

Asbtract ( )   PDF (2387KB) ( )  

References | Related Articles | Metrics

Mining area buildings are critical components of coal mining production. Dangerous deformations can severely threaten normal production and may even lead to safety incidents. In this study, the GNSS-InSAR fusion method is proposed to achieve high-precision deformation monitoring of mining area buildings. Taking the northern suburbs mining area of Xilinhot city,Inner Mongolia, as the study area, the dynamic deformation of mining buildings is obtained based on the GNSS,SBAS-InSAR, and GNSS-InSAR fusion methods with 30 scenes of Sentinel-1A image data and 35 sites GNSS data. The results show that the GNSS-InSAR fusion method is 43.9% more accurate than the SBAS-InSAR,which indicates that the proposed method provides better support for the deformation monitoring and safety assessment of mining area buildings. The combination of rainfall,temperature,and time-series deformation results inferred that temperature is the primary cause of building deformation,and the impact of surrounding mining activities on the buildings is negligible. Moreover,during the monitoring period, all deformation values of the mining area buildings are below the permissible deformation thresholds.The results indicate no dangerous deformations and confirm that the buildings can continue safe operations.

Unsupervised loss function for dense matching in deep learning

LIU Xiao, GUAN Kai, JIN Fei, RUI Jie, WANG Shuxiang, LIN Yuzhun, CHENG Chuanxiang

2024, 0(11):  133-139.  doi:10.13474/j.cnki.11-2246.2024.1123

Asbtract ( )   PDF (1861KB) ( )  

References | Related Articles | Metrics

With the advancement of deep learning, supervised dense matching networks have achieved remarkable progress. However, obtaining real annotations for dense matching is challenging and costly, making unsupervised deep learning-based methods the future trend. Recently, numerous loss functions have been proposed for unsupervised dense matching. However, their combinations are complex and effects remain unknown. Therefore, this study investigates unsupervised loss functions for dense matching, analyzes the accuracy and matching performance of various losses, and validates the effectiveness of combined applications. The results demonstrate that the appearance matching loss plays a pivotal role in achieving convergence in accuracy for unsupervised dense matching networks. Combining appearance matching loss with left-right disparity consistency loss facilitates accurate non and weak textured region matches. Then, adding relative smoothing loss can better adapt to dark environments.

Dense reconstruction and measurement of outdoor weakly textured metals

FANG Li, LI Songlin

2024, 0(11):  140-146.  doi:10.13474/j.cnki.11-2246.2024.1124

Asbtract ( )   PDF (2361KB) ( )  

References | Related Articles | Metrics

In response to the challenge of measuring outdoor metal containers, characterized by their large size, weak texture, and complex surrounding environments, a novel measurement method for outdoor weak-textured metals based on multi-view dense reconstruction is proposed. This method achieves dimensional measurement by obtaining a three-dimensional characterization of the weak-textured metal surface. Initially, the traditional incremental structure from motion (SfM) algorithm is improved by introducing a new weighted connected graph, a degree-based initial match pair selection method, and a new viewpoint selection method based on the distribution of visible points, thereby increasing the density of sparse reconstruction. Subsequently, to address the issue of high-quality reconstruction in weak-textured metal areas, a depth estimation method based on sparse priors and a multi-resolution multi-window depth estimation framework are proposed. Experimental results demonstrate that the method presented in this paper can resolve the measurement challenges associated with outdoor weak-textured metals, with an absolute measurement error of 0.52 mm over a distance of 970 mm.

A method for modeling multipath errors in the BeiDou system for large-scale reference station networks

CHEN Jing, WANG Yipeng, MAI Jiankai, LI Qi, ZHANG Yongfeng

2024, 0(11):  147-150.  doi:10.13474/j.cnki.11-2246.2024.1125

Asbtract ( )   PDF (1393KB) ( )  

References | Related Articles | Metrics

Multipath effects in global navigation satellite system (GNSS) observations are environment-dependent and cannot be eliminated or mitigated through differential methods. They significantly impact the accuracy of GNSS data processing and service performance. The BeiDou system comprises satellites in high, medium, and low Earth orbits. To mitigate the impact of orbit repetition differences among satellites, the multi-point hemispherical grid model (MHGM) based on spatial domain modeling presents an effective solution. However, due to the MHGM method's high number of parameters for estimation and its demand for substantial computational resources such as memory and CPU, it is generally unsuitable for modeling multipath errors in large-scale reference station networks. This paper investigates a partitioning modeling method based on public stations, achieving rapid and effective modeling of multipath errors in large-scale reference station networks. The effectiveness of this approach is validated using observation data from 221 reference stations of the BeiDou ground-based augmentation system in Guangdong province. Test results demonstrate a service performance enhancement of 11.2% for the existing system in Guangdong province, without requiring additional hardware installation. This research holds scientific and engineering significance in improving the accuracy and reliability of data processing services in the BeiDou system.

Application of multi-source point cloud fusion technology in urban renewal data acquisition

MU Cuiwei, WANG Zhaoze

2024, 0(11):  151-155.  doi:10.13474/j.cnki.11-2246.2024.1126

Asbtract ( )   PDF (1978KB) ( )  

References | Related Articles | Metrics

Multi-source point cloud fusion technology is playing an increasingly important role in urban renewal data collection. This technology can achieve comprehensive and accurate perception of urban features by integrating point cloud data obtained from different devices and sensors. In the process of urban renewal, the use of multi-source point cloud data fusion technology can quickly and efficiently obtain urban basic spatial data, providing rich spatial data information for urban renewal planning, design, and decision-making. Through the fusion processing and analysis of point cloud data obtained from data collection equipment such as vehicle-mounted mobile measurement systems, motorcycle-mounted mobile measurement systems, drone-mounted mobile measurement systems, and stationary 3D laser scanners, various basic spatial data required for urban renewal can be quickly obtained. Furthermore, an in-depth analysis and research on the accuracy of the data in plane coordinates is carried out, thereby promoting the popularization and application of this technology in urban renewal data collection, and providing beneficial assistance for accelerating the intelligent and fine development of urban renewal work.

Flood disaster risk and urban resilience assessment in Guangzhou based on open-source information and SDGSAT-1 nighttime light data

LIU Jincang, DONG Jing, WANG Huanhuan, Lü Mingyang, DING Yixing

2024, 0(11):  156-161.  doi:10.13474/j.cnki.11-2246.2024.1127

Asbtract ( )   PDF (1520KB) ( )  

References | Related Articles | Metrics

Guangzhou is one of the 31 key flood control cities in China, which has the characteristics of heavy rain flood, transit flood and heavy typhoon impact. Researches on flood disaster risk and urban resilience assessment can provide scientific references for building a resilient city, improving modern urban governance, and achieving sustainable development in Guangzhou. This paper uses open-source geo-information and statistical data, combined with the nighttime light data of SDGSAT-1, to construct 18 flood evaluation indicators. Combined with the analytic hierarchy process, a model for flood disaster risk assessment in Guangzhou is formed. In addition, this paper refines 18 indicators to evaluate the urban resilience of Guangzhou from 2013 to 2022. The comprehensive analysis results indicate that the northern mountainous area of Conghua, the vicinity of Liuxi River in Conghua-Huadu-Baiyun, the Dongjiang-Zengjiang area of Zengcheng, and the main urban area with concentrated population are relatively high-risk areas. It also indicates that the urban resilience has significantly improved. However, there is still room for improvement in terms of adjusting population structure and increasing public safety expenditures.

Fusion technology and application of GIS based marine digital system: taking Zhejiang Province's Intelligent Ocean Control as an example

ZHU Junxia, MAO Keqin, ZHANG Can

2024, 0(11):  162-166.  doi:10.13474/j.cnki.11-2246.2024.1128

Asbtract ( )   PDF (1371KB) ( )  

References | Related Articles | Metrics

Application scenarios are the main carrier of digital reform, serving as the backbone and link for achieving digital empowerment and institutional reshaping. Based on the pilot work of application scenarios in various counties (cities, districts), the use of fusion technology can avoid duplicate financial investment at all levels, reduce the burden of grassroots operation and maintenance, and further enhance the effectiveness of digital reform. This article focuses on four existing marine related business systems and application scenarios:marine spatial resource supervision, marine ecological warning and evaluation, marine disaster perception and prevention, and marine economic monitoring and evaluation. From the dimensions of business, technology, and interaction, GIS based digital fusion technology for marine related business is studied, achieving data fusion, technology fusion, and interaction fusion between business systems and application scenarios. Based on the fusion architecture, multiple applications of intelligent control of the ocean in Zhejiang province are integrated and tested.

Research on practical teaching and learning for industry-education integration of surveying and mapping geographic information specialties

SHI Guigang, ZHENG Runqiang, SHU Zhen

2024, 0(11):  167-171.  doi:10.13474/j.cnki.11-2246.2024.1129

Asbtract ( )   PDF (1275KB) ( )  

References | Related Articles | Metrics

Against the background of rapid development of economy and industrial structure,the requirements for practical ability of surveying,mapping and geographic information majors have increased significantly. Local undergraduate colleges and universities face the problems of insufficient resources and in-depth school-enterprise cooperation in practice teaching,which limits the cultivation of students' engineering practice ability. To cope with these challenges,this paper proposes to strengthen the integration of industry and education,optimize teaching resources,deepen the cooperation between schools and enterprises,improve the incentive of competition,improve the teaching links,innovate the teaching methods,and establish a quality assurance mechanism oriented to the results. Through the in-depth promotion of industry-education integration and the data-driven evaluation system,the content and methods of practical teaching are adjusted in time to improve the quality of practical teaching and the comprehensive quality of students. These measures aim to cultivate high-quality applied talents that meet the needs of modern industries,and to improve the educational quality and market competitiveness of local undergraduate colleges and universities.

Classification of tunnel point clouds based on improved cascaded BP neural network

DING Penghui, LI Zhiyuan, LIU Yi, WANG Zhenghui

2024, 0(11):  172-176.  doi:10.13474/j.cnki.11-2246.2024.1130

Asbtract ( )   PDF (1795KB) ( )  

References | Related Articles | Metrics

Efficient classification of tunnel point cloud data is crucial for safety monitoring and 3D reconstruction in underground transportation and mining operations, as it facilitates the comprehensive exploration and utilization of point cloud data. This study addresses issues in existing tunnel point cloud classification methods, such as noise sensitivity, low processing efficiency, and susceptibility to overfitting, by proposing a cascaded backpropagation (CBP) neural network classification method optimized with an early stopping mechanism and adaptive parameter tuning. Firstly, the Trimble RealWorks software is used to separate tunnel and ground point clouds. Then, local geometric features are extracted using spherical neighborhood space and covariance matrix eigenvalues to construct feature vectors. Finally, an improved CBP network is employed to hierarchically classify internal tunnel lighting equipment, signage, and various pipelines, thereby enhancing classification efficiency and accuracy. Experimental results demonstrate that the improved CBP neural network achieves high accuracy and reliability in tunnel point cloud classification, significantly improving data processing efficiency and providing data support for tunnel maintenance, renovation, and safety management.

Farmland soil moisture monitoring based on UAV multispectral imagery

ZHAO Guiping, XU Fajun

2024, 0(11):  177-182.  doi:10.13474/j.cnki.11-2246.2024.1131

Asbtract ( )   PDF (1953KB) ( )  

References | Related Articles | Metrics

Using drones to monitor soil moisture is low-cost, convenient, fast and accurate, and has important practical significance for intelligent management of farmland areas. This study selected Liangfeng Farm as the research area, where a drone equipped with a multispectral camera is used to monitor soil moisture. Through gray correlation screening, soil moisture sensitive spectral data are selected, and regression analysis was performed with the measured soil moisture data to construct a soil moisture inversion model based on UAV multispectral remote sensing. Through comparative analysis of the results of the NIR-RE-G model and the B-R-G-RE-NIR model, it is found that the determination coefficient R² is both greater than 0.77. The B-R-G-RE-NIR model is better than the NIR-RE-G model in terms of accuracy evaluation results of R² and RMSE, so the overall inversion results of both models have higher accuracy. Therefore, this study verified the effectiveness and feasibility of the NIR-RE-G model and the B-R-G-RE-NIR model in soil moisture monitoring in this region, which provides an effective method and reliable reference for rapid monitoring of soil moisture in large-scale farmland.