This PDF file contains the front matter associated with SPIE Proceedings Volume 13514, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

During the process of partitioning the input image into blocks, existing visual transformers damage the internal structural information of the image blocks, and the long-range attention mechanism easily ignores the local unique properties of the image, resulting in inferior performance to traditional convolutional networks. This study proposes a new neural network RTnet, which combines RepVGG network with Transformer to fully utilize the advantages of RepVGG network, the advantages of RepVGG network in feature extraction, while also introducing Transformer's self-attention mechanism to further capture cross spatial dependencies and improve the accuracy of change detection to a certain extent. Through experiments on the LEVRO-CD and DSIFN-CD datasets, it was through experiments on the LEVRO-CD and DSIFN-CD datasets, it was shown that the RTnet network achieved better Change Detection (CD) performance compared to previous similar architectures, and the accuracy of change detection increased by 1.41%. detection increased by 1.41%.

Compared to vertical wells, horizontal wells may exhibit identical logging curve characteristics corresponding to different target formations, and the same target formation can display varying curve features. To clarify these response character-istics, a numerical simulation was conducted using the fast Hankel transform method to simulate array induction logging responses for horizontal wells. The simulation covers both constant-inclination well trajectories in simple formations and waveform well trajectories in Oklahoma formations. Conduct a detailed analysis of the impact of well trajectory variations on array induction logging responses in horizontal wells.

As the demand for mineral raw materials continues to grow, the frequency of mining has increased year by year. Quickly and accurately identifying mining boundaries has become one of the core issues in mining management. Aiming at the problems of complex feature scenes and ambiguous boundary features in open pit mining areas, this paper proposes an innovative USF-Net method, which combines SAM with task-specific knowledge to effectively improve the recognition accuracy of open pit mining boundaries. By merging the generic knowledge from the SAM model with the specific requirements of the open pit mine extraction task, the method demonstrates excellent performance in complex mining scenarios. The experimental results show that USF-Net has high robustness and accuracy in open pit mining area boundary identification, which provides a new technical idea for open pit mining management.

Semantic segmentation of high-resolution remote sensing images is vital for various applications, including environmental monitoring, urban planning, and land resource management. Nonetheless, these images frequently pose challenges owing to substantial intra-class similarity and considerable inter-class variation, which often result in the misclassification of small target objects. To address these issues, we introduce an innovative deep learning network model that employs pixel memory sharing to improve feature representation and enhance segmentation accuracy. Our model employs the Swin Transformer as the backbone, which effectively reduces the computational burden associated with high-resolution feature extraction while maintaining rich contextual information. Despite these advancements, semantic information loss during the encoder phase poses a significant question, potentially compromising the overall segmentation accuracy. To address this question, we introduce the Pixel Memory Feature Fusion Module (PMFFM), which retains original pixel features and mitigates detailed information loss during feature propagation through effective memory sharing and fusion. PMFFM assigns different weights to each pixel to represent memory and guide multi-level feature fusion. During the decoder, the integration of feature memory sharing with PMFFM ensures the completeness of features at each layer, allowing each multi-scale feature layer to be representative and distinctive. Extensive experiments conducted on the ISPRS Vaihingen dataset demonstrate that our proposed method substantially enhances segmentation performance, leading to significant improvements across key evaluation metrics.

The remote sensing image segmentation of eucalyptus faces the problems of missed judgment and misjudgment due to the lack of local information caused by the complex backgrounds. In this study, a semantic segmentation method for extracting eucalyptus is proposed based on the Swin transformer network framework with the fusion of double-layer attention, which a dual attention mechanism, termed as ECA-SimAM, is constructed by the fusion of the ECA and SimAM attention modules. The SimAM attention module is introduced in the global average pooling process of the ECA attention mechanism, which can more accurately measure the difference between each feature and its mean value, and obtain more accurate attention weight values, paying more attention to the areas that play a decisive role in segmentation, and improving the ability to capture important feature information for extracting eucalyptus. A fused GF-1 image of the study area is selected as the data source to construct the eucalyptus semantic segmentation datasets, and the experiment of extracting eucalyptus is conducted by using the Swin Transformer with ECA-SimAM. Compared with Deeplabv3+, U-net, Swin Transformer and the Swin Transformer with ECA, the results show that the accuracies of the improved method in extracting eucalyptus is significantly improved, which the Accuracy, Recall, F1-Score and IoU reach the highest values of 90.52%, 96.63%, 93.48%, and 87.76%, respectively, and the edge segmentation is more detailed, resulting in the optimal segmentation results for extracting eucalyptus.

Shizhai mining area is located in the Nanling tungsten-tin polymetallic metallogenic belt, with frequent geological structure activities and strong magmatic activities, which provides superior geological conditions for mineralization. The aim of this study is to provide insight into the metallogenic conditions at deep within the mining area and to identify favorable metallogenic structures. To this end, we have adopted the Controlled Source Audio Magnetotelluric Method (CSAMT), which is known for its large exploration depth, high anti-interference capability and high lateral resolution. Through the application of CSAMT technology, we have successfully revealed the distribution characteristics of the major electrical layers in the mining area, depicted the distribution of fault structures in detail, and traced the extension of these structures to depth. The results of the study inferred seven potential fault structures. In addition, combined with the research results of geological and geochemical exploration, we constructed a comprehensive geological-geophysical-geochemical prospecting model for typical lead-zinc deposits in Yangshan County. The model will provide a solid geophysical foundation and technical support for future deep prospecting prediction research in this area.

In view of the serious land subsidence in Nanjing, this paper uses 33 Sentinel-1A orbit lifting data from January 2021 to. April 2023 based on PS-InSAR and SBAS-InSAR technologies to monitor the subsidence in the main urban area of Nanjing. The monitoring results showed that the uneven settlement in Nanjing was obvious during the study period, with the settlement rate ranging from -50. 5 mm/a to 26 mm/a, and the cumulative settlement between -262. 2 mm and 88. 6 mm, and the most serious settlement areas were located in Moling Street, Jiangning District, Dingshan Street, Pukou District, and Honghua Street, Qinhuai District. Research results indicate that there is a certain spatial correlation between the process of urbanization, ground fissures and faults, and ground subsidence.

To analyze the formation mechanism of an island in a lagoon-type bay, the coupling model of horizontal 2-D tidal current, sediment transport, and terrain evolution is adopted to reproduce the morphological features of Shuidong Bay and the focus is on the formation and evolution of Dazhou Island in Shuidong Bay. The calculation results are consistent with the actual topography. Two models of sediment transport, i.e., total-load and bed-load sand transport models, are considered. The simulation shows that Dazhou Island in Shuidong Bay is formed by the continuous deposition of silts in the lagoon under the influence of rising tide and is a geomorphic form of flood tide delta. The total-load sand transport model is more suitable for simulating the terrain evolution process than the bed-load sand transport model. The simulation results also replicated the morphological changes of the mid-bay island over a 25-year period. This study explores the formation mechanism and evolution process of the general bay morphology through numerical simulation.

As the resolution of Earth observation imagery advances, many of the downstream industries that rely on remotely sensed imagery can also advance. Due to the physical limitations of the optical sensors carried by satellites, the growth of the resolution of remote sensing images becomes difficult. Therefore, it is becoming increasingly important to boost the resolution of Earth observation images by methods other than upgrading the physical components, such as super-resolution. As a computer vision task, convolutional neural networks (CNNs) perform well on super-resolution tasks. Transformer-based models also show good performance on it. The new model proposed in this paper, Convolutional Transformer Generative Adversarial Network (CTGAN). It promotes image super-resolution by balancing local features with global features. Results on real satellite datasets demonstrate the effectiveness of the CTGAN model.

Global Navigation Satellite System Interferometric Reflection (GNSS-IR) has become a current research focus. Soil Moisture (SM) monitoring can be achieved using associated technologies, which have significant application potential. To further improve the prediction accuracy of inversion SM, this paper proposes a SSA-ELM model to retrieve SM. This model utilizes the feature parameters extracted from signal-to-noise ratio (SNR) observations as inputs for the model. Then, solve the hyperparameters of the ELM model by SSA. Finally, the performance of this model is compared with that of the BP Neural Network (BPNN) and ELM model. The Experimental results demonstrate that: the SM inversion for the PRN32 satellite in the SSA-ELM model has a determination coefficient of 0.932, RMSE of 0.024, and MAE of 0.019, all are the optimal values in the experiment. The inversion results are more stable and can better reflect the changes of SM, demonstrating the reliability of this model.

Soil organic matter is an important index to evaluate soil fertility. Remote sensing technology has become a fast and cheap method to estimate soil organic carbon. However, we ignore the spatial heterogeneity and scale effect of the relationship between remote sensing parameters and soil organic matter, which is inconsistent with the reality. MGWR can effectively reveal the spatial non-stationarity and optimal scale of the relationship between variables. In this paper, the SOM content of 111 sampling points in Shouguang County, Weifang City and Changyi City was used as the dependent variable and 17 environmental covariables such as elevation, slope and slope direction were used as independent variables to construct the MGWR remote sensing model for soil organic carbon estimation. The predicted results were compared with the OLS and GWR models. The results show that: (1) the goodness of fit of MGWR is better than GWR, and the goodness of fit of GWR is better than OLS. (2) SPI, land use type, DVI, RVI, NDVI, NDII, topographic relief and slope had the largest effect scale on soil organic matter concentration (109); Followed by mndwi (107) and air temperature (105); The effect scale of ground temperature, TWI, soil type and precipitation is relatively small, the value is in the range of 90 to 100; The effect scale of slope side is the most limited, only 50. (3) The spatial non-stationarity of MGWR and GWR is not obvious, and they have strong spatial homogeneity.

Compared with traditional deformation monitoring methods, GBSAR has the advantages of all-weather, near real-time and high spatial-temporal resolution, which provides an effective deformation monitoring method for local region. In this paper, the slope of mining area is continuously monitored by using GBSAR during the treatment period. The results of time series show that the maximum cumulative displacement is 3.8 mm at the same time, in order to verify the accuracy of GBSAR, monitoring experiments are carried out. The maximum difference between the actual displacement value and the monitoring value was -0.37mm, the minimum was -0.01mm, and the average root mean square error was 0.2mm. TS30 total station is used as auxiliary data to evaluate the external coincidence accuracy of GBSAR monitoring results. The results show that the monitoring results of GBSAR and total station are highly consistent, and the correlation of deformation monitoring points is 0.98 and 0.93, respectively, indicating the reliability of GBSAR monitoring technology in slope monitoring. Finally, through field investigation and analysis, it is concluded that the deformation of the slope is related to artificial irrigation, and the slope remains basically stable. The results show that GBSAR has the characteristics of high spatial resolution and high precision, which can continuously monitor small areas at sub-millimeter level and provide necessary technical support for disaster reduction and prevention.

To address the issues of missing, misclassified, and distorted building features in the U-Net's single-channel double convolution and skip connection approach, this paper proposes a DC-CTransNet building extraction model. This model expands the receptive field by introducing a DC-Block module with dual-channel triple convolutions and feature concatenation. Additionally, it replaces the traditional skip connection with a CTrans component that includes a Channel Cross-Fusion Transformer (CCT) module and a Channel Cross-Fusion Attention (CCA) block to enhance building detection performance. Experimental results show that DC-CTransNet achieves OA, MIoU, Precision, and FWIoU scores of 98.5%, 97.5%, 95%, and 97.1%, and 94.3%, 96.9%, 96.1%, and 95.4% on the Aerial Image Dataset and Aerial Imagery Labeling datasets, respectively. The building extraction accuracy improved by 1.8%, 2.1%, 4.1%, 3.2%, and 4.1%, 2.8%, 4.2%, 2.0%, respectively. This research provides valuable insights for similar models in building extraction.

Random Error Code (REC) frequently occurs during the sending and receiving of high-speed data among multiple devices in remote sensing satellites, which brings great challenges to the safety of space equipment and the application of remote sensing data. To address this issue, this paper proposes a low-cost solution, which can realize the demand of real-time error-free transmission of 100Gbps rate among the satellite-mounted data processing devices under the condition that the occupancy rate of the hardware resources (computational resources and caching resources) is not higher than 3% by the measures of data segmentation, parallel processing, and error correcting coding.

The integration of remote sensing technology and deep learning has significantly advanced object recognition in high-resolution satellite imagery. However, most existing datasets mainly focus on stumpy objects, while slender targets, characterized by a height that significantly exceeds their length or width, have been overlooked. This gap highlights the need for specialized datasets to improve recognition of slender objects. In response, this study developed the Transmission Tower Oriented Bounding Box (TT-OBB) dataset, which systematically classifies transmission towers as representative slender targets into four distinct categories. To validate the effectiveness of deep learning methods in recognizing these structures, comprehensive experiments were conducted using several mainstream detectors on the TT-OBB dataset. The results demonstrate the practicality of this dataset, highlighting its utility as a reference for deep learning applications in transmission tower identification, and providing valuable insights for model selection in real world scenarios

In order to improve the accuracy of airborne LiDAR point cloud filtering algorithm in terrain complex regions, an adaptive cloth simulation filtering algorithm based on terrain roughness is proposed. Firstly, the point cloud data is preprocessed to extract three terrain features: Gaussian curvature, average curvature, and slope. Principal component analysis is used to linearly combine them into a comprehensive terrain roughness. Different grid scales are divided according to the terrain, and different stages are set based on the roughness of each subregion. Different fabric hardness, fabric resolution, and ground point thresholds are set. Finally, fabric simulation filtering is performed. Three sets of data publicly released by the International Society for Photogrammetry and Remote Sensing (ISPRS) are used to validate the algorithm, and the experimental results show that, compared with the traditional cloth simulation filtering algorithm, the class I error, class II error and total error are reduced to achieve a better filtering effect.

Typhoons, originating from tropical ocean surfaces, are among the most severe natural disasters worldwide, often leading to significant loss of life and property. Research indicates that the Total Electron Content (TEC) of the ionosphere experiences various disturbances before and after a typhoon event. This study utilizes global ionospheric data provided by the Chinese Academy of Sciences (CAS) to analyze the anomalies in TEC during a 15-day period surrounding Super Typhoon "Meranti" (No. 14) in 2016, employing Singular Spectrum Analysis (SSA) and the Sliding Interquartile Range (SIQR) method to detect perturbations. The results show that the anomalies detected by the SSA method are smaller than those detected by the SIQR method, but the number of detected abnormal periods increases significantly, indicating that the SSA method is more sensitive to anomaly detection of ionospheric TEC. Taken together, the SSA method shows significant advantages in detecting ionospheric TEC disturbances along the typhoon path. It can identify abnormal periods more comprehensively, highlight the abnormal characteristics of the ionosphere, and provides an effective tool for understanding the impact of typhoons on the ionosphere.

During magnetic storms, scintillation is instigated by significant disturbances in the ionosphere that deviate from the mean level. Utilizing global ionospheric TEC and ROTI data, this study analyzes the global ionospheric impact during the substantial magnetic storm of December 2015. The analysis reveals that during the magnetic storm, the southern hemisphere ionosphere was primarily characterized by positive-phase storms, whereas the northern hemisphere exhibited brief positive-phase storms succeeded by prolonged and intense negative-phase storms. The response to magnetic storms is more pronounced at low to mid-latitudes compared to high latitudes, with ionospheric irregularities in the equatorial region being particularly significant. Analysis of changes in the high-latitude ionospheric ROTI map for the northern hemisphere during the magnetic storm indicates substantial ionospheric irregularities at high-latitude poles, characterized by large elliptical structures, suggesting the occurrence of a significant ionospheric scintillation event. The findings indicate pronounced temporal and spatial differences in ionospheric variations throughout this magnetic storm.

The ionosphere is characterized by a high concentration of ionized particles, plays a critical role in the propagation characteristics of radio signals in the upper atmosphere. During major geological events, such as earthquakes and volcanic eruptions, the ionosphere often exhibits anomalous disturbances. In the event of an earthquake, energy accumulated within the Earth’s crust is released and propagates along fault lines, generating waves—including Rayleigh waves, acoustic waves, and gravity waves—that ascend into the ionosphere and induce fluctuations in Total Electron Content (TEC). This study investigates ionospheric disturbances associated with the magnitude 7.8 earthquake that impacted the Kaikōura region of New Zealand on November 13, 2016. Drawing on observational data from the Geo-Net Data Center, derived from New Zealand’s network of Continuously Operating Reference Stations (CORS), this research examines ionospheric TEC and its rate of change (dTEC) in response to the earthquake. Two-dimensional dTEC distribution maps and distance-time plots are generated to evaluate the spatiotemporal characteristics and propagation dynamics of ionospheric disturbances during the event. By analyzing the propagation characteristics of ionospheric disturbances across multiple time intervals, this study identifies the presence and unique propagation patterns of various seismic wave phenomena transmitted through the ionosphere. The results contribute to a deeper understanding of the mechanisms underlying earthquake-induced ionospheric disturbances and present new insights and methodologies for monitoring and predicting such events.

To address the high computational load and resource consumption associated with traditional carrier phase fixing methods in short baseline positioning, this paper proposes a novel approach that utilizes raw dual-frequency high-precision satellite navigation carrier phase inter-station differential information to eliminate the effects of inter-station clock errors. This method is based on Dual-Frequency AFM (Ambiguity Function Method) for fixing the integer ambiguities of short baselines in a single epoch and single satellite scenario. By improving the AFM objective function, the method directly employs the inter-station differential information of dual-frequency carrier phases without relying on previous epoch caching or pseudorange information, effectively reducing computational requirements. Using a set of measured short baseline dual-station observation data, we validate the stability of the inter-station single-differential carrier phase ambiguity compensation values and their differences from true values. The results indicate that under an ideal condition with a 0.5m baseline, the error rate is less than 1%, and the computation speed increases by nearly 99%. This method effectively facilitates the fixing of inter-station single-differential carrier phase ambiguities under dual-frequency observable conditions and demonstrates promising application prospects.

The purpose of this paper is to construct a precise management and risk assessment system for mariculture by integrating Geographic Information System (GIS) and remote sensing technology. By integrating the spatial analysis capability of GIS and the real-time monitoring function of remote sensing technology, it realizes the comprehensive perception and precise management of mariculture environment. The system includes several modules such as data collection and processing, precise management strategy formulation, risk assessment and early warning, aiming to improve the scientific and intelligent level of mariculture. Taking a case study of a sea area in Fujian as a case study, the effectiveness of the system in improving aquaculture efficiency and reducing risks is verified, which provides a strong support for the sustainable development of mariculture in the sea area.

Due to the different load demands in different areas of the city, the effect of trunk line planning is difficult to ensure the effective utilization of electric energy. Therefore, a study on urban distribution network trunk line planning based on high-resolution remote sensing images is proposed. Considering the objective demand of the load, the distribution network is divided into a three-layer grid model according to the three voltage levels of the distribution network: high voltage, medium voltage and low voltage, combined with the high-resolution remote sensing image information. In the specific division stage, the IBIWO algorithm is introduced. After the initial line architecture is divided based on the three-layer grid model, the target function value of the line planning is set in combination with the power flow, and with the help of the competitive exclusion mechanism, Determine the optimal architecture for the space diffusion phase. In the test results, the overall active load of the grid in a single day reached 1108.22kW, which was 34.92kW and 26.41kW higher than that of the control group, respectively. The proportion of reactive output in different load periods was only 5.24%, which had obvious advantages compared with the control group.

For satellite, airborne platform detection and communication systems, according to its typical working characteristics, the mathematical models of the motion characteristics of satellite and airborne platform are established, the error factors that affect the measurement accuracy of satellite and airborne sensor and the target detection capability of satellite and airborne platform are analyzed. Based on the modular design idea, a universal digital prototype of satellite airborne cooperative detection platform with adjustable parameters has been formed. Analysis results show the digital prototype of satellite airborne cooperative detection platform can realize the function and performance index of the physical prototype, respond effectively to changes in input parameters, provide a reliable research method for the index demonstration and analysis of physical prototype and the optimal design of scheme, and it will be widely used in the future.

With the rapid development of drone technology, drones are being used more and more widely in fields such as logistics and distribution, agricultural monitoring, disaster relief and environmental protection. However, path planning in dynamic environments remains one of the major challenges for UAVs. Traditional path planning methods such as Dijkstra and A* algorithms perform well in static environments but have limited effectiveness in dynamic environments. This paper proposes an UAV path planning method based on Ant Colony Optimization (ACO) optimization. By introducing a dynamic pheromone update mechanism, environmental prediction technology, and an adaptive parameter adjustment strategy, the adaptability and efficiency of the algorithm in dynamic environments are improved. Simulation experiments have shown that the improved ACO algorithm is superior to the traditional ACO and A* algorithms in terms of path length, calculation time, success rate and path smoothness. The research in this paper provides a feasible solution for the autonomous flight of UAVs in dynamic environments and provides new ideas for the expansion of intelligent optimization algorithms in practical applications.

In the backdrop of swift urban expansion, as urban areas rapidly develop, the management of land designated for construction has emerged as the principal method for managing urban territory. However, this process has led to inefficient utilization of building land, resulting in a considerable quantity of underutilized land that hinders the city's future growth. Consequently, an evaluation index system tailored to local conditions has been established for identifying inefficiently used urban land. This system utilizes multisource statistics to facilitate spatial intersection assessment regarding building land The results indicate that the identification accuracy of inefficiently used land in the study area is 85.265%, this inefficiently used land is primarily concentrated in the central, western, and northern areas within the city. Currently, the city is in the later stages of urbanization, characterized by a slower pace of urbanization but a rising level of economic development.

GNSS reflectometry is a low cost, high spatial/temporal resolution microwave remote sensing method for estimating wind speed. The Delay Doppler Maps (DDMs) obtained from the GNSS-R observations show the sea surface roughness at the point of reflection and nearby areas. We propose a model that combines Transformer and Kan, which can use DDMs to predict wind speed. The DDM data features are extracted using the self-attention mechanism of Transformer, then the wind speed estimation is achieved by MLP, finally we correct the distribution of the retrieved wind speeds using Kan. The RMSE for this method can reach 1.88 m/s. The biases in the low, medium and high wind speed ranges are reduced by 50.5%, 50.0% and 29.5% respectively, indicating that the distribution of KAN-corrected wind speeds is more similar to the reference wind speed.

Global Navigation Satellite System-Reflectometry (GNSS-R) technology has been widely applied in soil moisture retrieval due to its low cost, short revisit periods, and precise positioning. To fully exploit global contextual information and long-range dependencies in Delay-Doppler maps (DDMs) from the Cyclone Global Navigation Satellite System (CYGNSS) and related auxiliary data, this paper proposes a novel method—Cross-Modal Transformer Network (CT-NET). CT-NET uniquely combines Transformer's multi-head self-attention and cross-attention mechanisms with a complementary fusion. The self-attention mechanism captures intrinsic features of GNSS-R multimodal data, while cross-attention enhances interconnectivity between different modal data. The Yellow River Delta is selected as the experimental area to validate the proposed method using soil moisture observations from the Soil Moisture Active Passive (SMAP) satellite. Compared with three state-of-the-art models, experimental results show that CT-NET outperforms other methods, confirming the innovation and superiority of this approach for soil moisture retrieval.

The measurement criteria and evaluation methods of radar seeker anti-jamming performance evaluation are a complex problem, and there are many methods for performance evaluation, and appropriate evaluation methods need to be adopted for different targets. In this paper, the evaluation results are obtained by establishing a binary tree evaluation model based on statistical decision-making, establishing an index system and judging a large number of experimental data, and finally confirming the feasibility of the method by comparing it with the comprehensive evaluation method based on fuzzy level.

The images from the linear array sensors circular scanning perpendicular to orbit have the properties of super wide breadth and high resolution, which also comes with the problem of stereo matching. One method to improve the reliability and efficiency of stereo matching is to search the same one point in linear array. Refer to epipolarity models of linear pushbroom image and plane array remote sensing image, we construct the epipolarity model of the linear array circular scanning image based on a strict imaging model. In this work, we obtain the epiplolarity of the images captured by the linear array sensors circular scanning perpendicular to the orbit using these models. The result shows that the epipolarity of the circular scanning image presents different shapes in different region, including most part with line, a few with parabola and hyperbola in the boundary. These conclusions could provide the progress of three dimension reconstruction.

Deep learning-based methods for real-time small target detection are applicable to various industrial tasks, such as real-time traffic monitoring, military reconnaissance, and marine transportation. However, the high-precision detection of small objects in images with a complex background remains a challenge due to insufficient feature representation and background confusion. To address this challenge, this study proposes an efficient detector called multi-scale Feature Enhancement and Background Denoising YOLO (FEBD-YOLO), which includes a multiscale Feature Extraction Enhancement Module (FEEM) for effective feature extraction, a Hybrid Local Channel Attention module (HLCA) for enhancing foreground small target features, and a Background Denoising Module (BDM) for eliminating noise interference. The enhancement of the three modules in the feature extraction and feature fusion sessions provides the FEBD-YOLO with a high-precision detection capability for multi-scale small targets in images with a complex background. The proposed method can reach an accuracy of 0.782 (in terms of mAP@0.5) on the Martian crater datasets and 0.751 on the VEDAI datasets, surpassing several benchmark models and the other state-of-the-art models. In addition, the robustness of the FEBD-YOLO is verified under different simulated degradation conditions, which expected to be applied to satellite-based real-time detection tasks.

Previous studies on crop classification methods based on deep learning for multitemporal images had already determined the number of inputs multi temporal images in the network structure design stage. However, in reality, due to satellite revisit cycles, weather, and other reasons, stable and clear remote sensing images (RSIs) cannot be continuously obtained. Once a period of image is missing from the multi temporal image sequence, the entire method cannot be used. Although methods such as interpolation and using other images instead can be used to address this issue, they greatly reduce the classification accuracy and stability of the methods, limiting their large-scale application. In response to the above issues, we first proposed a flexible multi temporal RSI dataset. For this dataset, an improved version UNet is constructed to train the model. Crop classification experiments shows that this model can be used without limiting the number of RSI periods and time inputs, and the classification accuracy gradually increases with the increase of image periods.

This research delves into the uses of big data, in public health and emphasizes its crucial role in assessing health risks, data control of disease transmission and allocation of public health resources. Geological big data, serving as an resource could shed light on the complex connection between environmental factors and public health introducing new predictive models for health risks evidence-based strategies, for mitigating diseases and sensible distribution of public health resources. Facing challenges related to data accuracy and privacy protection as technical processing issues doesn't diminish the undeniable value of using this technology for making well informed decisions, in public health matters. Looking ahead to the synergy of advancements and regulatory improvements in disciplines and anticipating the integration of extensive geological data, into public health initiatives promises to enhance the precision and effectiveness of managing healthcare systems in the future.

TaiJing-4(03) satellite is the world's first high resolution flat-panel Ku-band Synthetic Aperture Radar (SAR) commercial satellite, characterized by the high resolution, light weight design, low cost, multiple imaging modes and long-term operation. A series of key technologies and innovations are adopted, such as lightweight and high-resolution Ku-band SAR payload, on-board intelligent imaging processing, non-uniform pulse sparse reconstruction of high resolution and wide width SAR imaging technology, target detection and recognition capability improvement, high precision and high stability attitude measurement and control, 2D guidance composite control, and multi-target imaging efficiency improvement. It achieves high resolution, high image quality and high image acquisition efficiency. The Ku-band SAR satellite has stronger perception of target details, and has multiple workmodes such as spotlight, sliding spotlight, strip, scanning, wide-width scanning, and multi-target imaging, which significantly improves satellite observation performance, coverage capability and application efficiency. It has important applications in high-resolution observation, emergency disaster relief, marine environment monitoring, resource investigation, etc.

The mobile measurement system has promoted the development of photogrammetry, drone mapping, and 3D laser scanning technology and has made breakthrough progress in supporting software development and professional applications. As an emerging surveying method, mobile surveying systems have been widely used in terrain surveying. In response to the advantages of fast information acquisition speed, comprehensive acquisition of ground features, and high data accuracy in mobile measurement technology, it is proposed to use unmanned aerial vehicles equipped with laser scanning systems and aerial digital cameras to collect data in coastal areas and obtain three-dimensional data information of ground features through post-processing. At the same time, combined with quality inspection work, the system analyzed the efficiency and accuracy of terrain surveying using mobile measurement systems; this provides a new technical method for large-scale terrain surveying and mapping.

The coordinated operation of UAV Swarm enhances positioning accuracy, reduces individual errors, and improves system robustness and real-time performance. Current research on cooperative positioning mainly focuses on multi-sensor fusion and technological optimizations. This article aims to find a high-precision and stable cooperative positioning solution for UAV Swarm by addressing the limitations of individual positioning coordination. It reviews existing research, analyzing the advantages and key technologies of multi-sensor fusion specific to UAV Swarm. The study also explores advancements in visual positioning optimization, signal anti-interference, and fusion filtering techniques to improve accuracy, offering valuable insights for the development of multi-UAV cooperative positioning systems.

A major challenge in the field of Hyperspectral Image (HSI) analysis is how to effectively use spatial-spectral information to improve land cover classification accuracy when labeled training samples are limited. In this paper, a novel HSI classification method termed Multiscale Neighborhood Structure Tensor Representation Classification (MNSTRC) was proposed. At first, a tensor-based neighborhood structure representation model is designed for HSI, which can reveal the representation relationships between tensor samples from different feature dimensions. Then, it adopts the minimum representation error of different orders as the overall model error, which can enhance the model's representation capability for tensor samples. Finally, based on the scale variability of tensor samples, the classification results of different scales are integrated to further improve the classification performance. The performance of MNSTRC is evaluated on two hyperspectral datasets, and the experimental results show that the proposed MNSTRC method can obtain higher classification accuracies when compared with some state-of-the-art methods.

The compass is regarded as one of the four great inventions in ancient China. While the traditional combination of magnetic needle and compass can provide orientation indications, it is bulky, inconvenient to carry, and has limited accuracy. This paper presents a digital compass based on the HMC5883L magnetoresistive sensor, designed for the precise acquisition of geomagnetic data. The data collected by the sensor is transferred to the STC89C52 microcontroller via the I²C interface, whereupon the microcontroller performs the calculation of the azimuth angle and displays the result on the LCD1602 display. Moreover, the system is capable of communicating with a computer via a USB-to-TTL module, thus facilitating additional processing and data display. Practical tests have demonstrated that the system is able to provide accurate azimuth information in real time with high positioning accuracy and stability, thereby offering a new supporting technology for digital guides.

Textile and costume intangible cultural heritage items carries significant social connotations and historical data value. The spatial distribution pattern and formation mechanism of textile and costume intangible cultural heritage items are investigated in this paper using ArcGIS spatial statistics and analysis tools. The methods used include nearest neighbour analysis, Lorentz curve, concentration index, kernel density estimation and geodetector. textile and costume intangible cultural heritage items show a concentrated distribution pattern, but the east-west distribution is unbalanced. Socioeconomic factors and historical factors influence the distribution of textile and costume intangible cultural heritage items. Through the analysis and integration of textile and costume intangible cultural heritage items, combined with ArcGIS geographic analysis software, geographical calculation and spatial expression, are carried out to excavate potential spatial information.