Remote Sensing Imagery Research Articles

Learning to Holistically Detect Bridges From Large-Size VHR Remote Sensing Imagery.

Bridge detection in remote sensing images (RSIs) plays a crucial role in various applications, but it poses unique challenges compared to the detection of other objects. In RSIs, bridges exhibit considerable variations in terms of their spatial scales and aspect ratios. Therefore, to ensure the visibility and integrity of bridges, it is essential to perform holistic bridge detection in large-size very-high-resolution (VHR) RSIs. However, the lack of datasets with large-size VHR RSIs limits the deep learning algorithms' performance on bridge detection. Due to the limitation of GPU memory in tackling large-size images, deep learning-based object detection methods commonly adopt the cropping strategy, which inevitably results in label fragmentation and discontinuous prediction. To ameliorate the scarcity of datasets, this paper proposes a large-scale dataset named GLH-Bridge comprising 6,000 VHR RSIs sampled from diverse geographic locations across the globe. These images encompass a wide range of sizes, varying from 2,048 × 2,048 to 16,384 × 16,384 pixels, and collectively feature 59,737 bridges. These bridges span diverse backgrounds, and each of them has been manually annotated, using both an oriented bounding box (OBB) and a horizontal bounding box (HBB). Furthermore, we present an efficient network for holistic bridge detection (HBD-Net) in large-size RSIs. The HBD-Net presents a separate detector-based feature fusion (SDFF) architecture and is optimized via a shape-sensitive sample re-weighting (SSRW) strategy. The SDFF architecture performs inter-layer feature fusion (IFF) to incorporate multi-scale context in the dynamic image pyramid (DIP) of the large-size image, and the SSRW strategy is employed to ensure an equitable balance in the regression weight of bridges with various aspect ratios. Based on the proposed GLH-Bridge dataset, we establish a bridge detection benchmark including the OBB and HBB tasks, and validate the effectiveness of the proposed HBD-Net. Additionally, cross-dataset generalization experiments on two publicly available datasets illustrate the strong generalization capability of the GLH-Bridge dataset.

Urban expansion and agricultural land loss: a GIS-Based analysis and policy implications in Hawassa city, Ethiopia

This study investigated the historical and future trends of urban expansion and its subsequent impact on agricultural land-use in Hawassa city, Ethiopia. A time-series of remote-sensing imageries from Landsat Thematic Mapper for the years 1984, 1990, 2000, and 2010 and Operational Land Imager for 2021 were used to extract the LULC information from the study area. Seven major land-cover classes’ waterbody, built-up, agricultural land, wetland, grassland, woody vegetation, and agroforestry were identified with visual image interpretation along with supervised image classification techniques using the maximum-likelihood algorithm for the study years. The urban and agricultural lands were then extracted from the original LULC data to quantify the extent, rates, and number of area conversion between the two. The Land Change Modeler module of TerrSet software was used to predict the spatial extents of built-up and agricultural lands in 2030 and 2050. The results showed that there have been significant changes between the LULC types in Hawassa city within the past 37 years, from which built-up and agricultural land have shown the most prevalent changes. It showed that built-up land has increased from 584.73 ha in 1,984–3,939.03 ha in 2021, representing a 573.65% increase at an annual growth rate of 15.50%. However, agricultural land decreased from 8,324.64 ha to 3,595.68 ha in the respective years, with a 56.81% decrease at a rate of −1.54% each year. A total of 3,148.74 ha (37.82%) of agricultural land was converted into built-up land within the past 37 years (85.10 ha per year, a rate of 1.02%. The built-up land is projected to increase to 5,009.85 ha and 6,794.73 ha from 2021 to 2030 and 2050, with annual growth rates of 3.02% and 2.50%, respectively. In the same years, agricultural land will decrease to 2,849.58 ha and 2033.46 ha by 2.31% and 1.50% annually, respectively, from which 64.76 ha (1.80%) and 48.41 ha (1.35%) will be converted into built-up land, respectively. Future planning and development in the city should consider the rapid increase in built-up land toward agricultural land areas and develop appropriate adaptation mechanisms for the local community, which is highly dependent on agriculture.

Geological Structures in The Formation of The Kendal Plains: Insights from Remote Sensing Imagery

Geological Structures in The Formation of The Kendal Plains: Insights from Remote Sensing Imagery

Assessing Land Use and Cover Changes arising from the 2022 water crisis in Southeast China: A comparative analysis of Remote Sensing Imagery classifications and Machine Learning algorithms

Abstract. The water crisis in the southeast region of China in 2022, caused by one of the worst heatwaves on record, was characterized by severe shortages of water resources, leading to challenges for local communities, agriculture, and industry. To analyze changes in land use and land cover (LULC) in the Jialing River region, Chongqing, China, we compared Remote Sensing (RS) imagery classifications before and after the intense heat waves of 2022. We evaluated the performance of two machine learning algorithms, KDTree KNN and Random Forest (RF), in LULC classifications. The classifications were carried out based on the RS images from the OLI/Landsat 8 system, NDWI index, and SRTM data. The model performances were similar, the classification accuracy showed that the RF algorithm was superior to KDTree KNN. The RF LULC classification and area calculation corroborate with the visual analysis, reaffirming the superiority of RF, which shows a decrease in water surface area, unlike DKTree KNN.

Ecophysiological variables retrieval and early stress detection: insights from a synthetic spatial scaling exercise

ABSTRACT The ability to access physiologically driven signals, such as surface temperature, photochemical reflectance index (PRI), and sun-induced chlorophyll fluorescence (SIF), through remote sensing (RS) are exciting developments for vegetation studies. Accessing this ecophysiological information requires considering processes operating at scales from the top-of-the-canopy to the photosystems, adding complexity compared to reflectance index-based approaches. To investigate the maturity and knowledge of the growing RS community in this area, COST Action CA17134 SENSECO organized a Spatial Scaling Challenge (SSC). Challenge participants were asked to retrieve four key ecophysiological variables for a field each of maize and wheat from a simulated field campaign: leaf area index (LAI), leaf chlorophyll content (C ab), maximum carboxylation rate (V cmax,25), and non-photochemical quenching (NPQ). The simulated campaign data included hyperspectral optical, thermal and SIF imagery, together with ground sampling of the four variables. Non-parametric methods that combined multiple spectral domains and field measurements were used most often, thereby indirectly performing the top-of-the-canopy to photosystem scaling. LAI and C ab were reliably retrieved in most cases, whereas V cmax,25 and NPQ were less accurately estimated and demanded information ancillary to RS imagery. The factors considered least by participants were the biophysical and physiological canopy vertical profiles, the spatial mismatch between RS sensors, the temporal mismatch between field sampling and RS acquisition, and measurement uncertainty. Furthermore, few participants developed NPQ maps into stress maps or provided a deeper analysis of their parameter retrievals. The SSC shows that, despite advances in statistical and physically based models, the vegetation RS community should improve how field and RS data are integrated and scaled in space and time. We expect this work will guide newcomers and support robust advances in this research field.

Exploiting Remote Sensing Imagery for Vehicle Detection and Classification using Robust Competitive Algorithm with Deep Convolutional Neural Network

Exploiting Remote Sensing Imagery for Vehicle Detection and Classification using Robust Competitive Algorithm with Deep Convolutional Neural Network

Enhanced semantic-positional feature fusion network via diverse pre-trained encoders for remote sensing image water-body segmentation

ABSTRACT In the era of increasingly advanced Earth Observation (EO) technologies, extracting pertinent information (such as water-bodies) from the Earth’s surface has become a crucial task. Deep Learning, especially via pre-trained models, currently offers a highly promising approach for the semantic segmentation of Remote Sensing Imagery (RSI). However, effectively adapting these pre-trained models to RSI tasks remains challenging. Typically, these models undergo fine-tuning for specialized tasks, involving modifications to their parameters or structure of the original architecture, which may impact their inherent generalization capabilities. Furthermore, robust pre-trained models on nature images are not specifically designed for RSI, presenting challenges in their direct application to RSI tasks. To alleviate these problems, our study introduces a light-weight Enhanced Semantic-positional Feature Fusion Network (ESFFNet), leveraging diverse pre-trained image encoders alongside extensive EO data. The proposed method begins by leveraging pre-trained encoders, specifically Vision Transformer (ViT)-based and Convolutional Neural Network (CNN)-based models, to extract deep semantic and precise positional features respectively, without additional training. Following this, we introduce the Enhanced Semantic-positional Feature Fusion Module (ESFFM). This module adeptly merges semantic features derived from the ViT-based encoder with spatial features extracted from the CNN-based encoder. Such integration is realized via multi-scale feature fusion, local and long-distance feature integration, and dense connectivity strategies, leading to a robust feature representation. Finally, the Primary Segmentation-guided Fine Extraction Module (PSFEM) further bolsters the precision of remote sensing image segmentation. Collectively, these two modules constitute our light-weight decoder, with a parameter size of less than 4 M. Our approach is evaluated on two distinct water-body datasets, indicating superiority over other leading segmentation techniques. In addition, our method also demonstrates exemplary efficacy in diverse remote sensing segmentation tasks, such as building extraction and land cover classification. The source codes will be available at https://github.com/zhilyzhang/ESFFNet.

Introducing a new index for flood mapping using Sentinel-2 imagery (SFMI)

Accurate surface water detection and mapping using Remote Sensing (RS) imagery is crucial for effective water and flood management and for supporting natural ecosystems and human development. In recent years, RS technology and satellite image processing have significantly advanced in flood and permanent water extraction, particularly in water index, clustering, classification, and sub-pixel analysis. Water-index-based techniques, distinguished by their quickness and convenience, offer notable advantages. The dynamic and extensive nature of surface water and flooded areas make the water index particularly effective for monitoring large areas. However, challenges arise due to the complexity of ground surfaces in aquatic environments, including shadows in built-up, vegetated, and mountainous regions, narrow water bodies, and muddy water. This research presents a new Flood Mapping Index using Sentinel-2 imagery (SFMI) designed to address these challenges and identify water and flooded areas more accurately. The SFMI utilizes visible and near-infrared bands derived from Sentinel-2 data, employing 10-m bands to compensate for errors arising from spectral and spatial changes more effectively. The SFMI index is designed based on the spectral signatures of various land cover classes, utilizing the potential of 10-m resolution bands to identify water bodies and flood areas. Unlike the most conventional methods, the SFMI identifies and extracts water and flood regions without complex thresholding, and thus mitigates the impact of irrelevant features, such as dense vegetation and rugged topography on the flood and water body maps. The proposed index was tested in two large areas with high spectral diversity, yielding promising results. The SFMI index demonstrates an average overall accuracy of 97.1% for pre-flood water extraction, 97.95% for post-flood water extraction, and 98% for flooded area extraction. Moreover, the results showed an average kappa coefficient of 0.958 for pre-flood water extraction, 0.965 for post-flood water extraction, and 0.978 for flooded area extraction. The performance of the SFMI index for extracting flooded areas (ΔSFMI) is superior to its performance for water extraction both before and after the flood. However, it is essential to note that the accuracy of the flooded area map is contingent on the accuracy of the water area map both before and after the flood. Thus, the SFMI index based on 10-m Sentinel-2 imagery accurately detects floods and water bodies over time, without relying on thresholding, making it suitable for flood management and monitoring various water bodies like dams, lakes, wetlands, and rivers. The findings underscore the applicability of the proposed SFMI index in diverse and spectrally rich areas, demonstrating its effectiveness in monitoring various surface water bodies, detecting floods, and managing flood crises.

Multi-Source Fusion Enhanced Feature Segmentation in Remote Sensing Imagery

Abstract. With deepening application of deep learning technology in the field of remote sensing, several challenges persist in the segmentation of remote sensing optical images. These challenges include: (1) insufficient availability of deep learning-based remote sensing semantic segmentation datasets; (2) inadequate utilization of multi-source remote sensing data in the field of semantic segmentation; (3) limited sample size for effective model training, as well as the need to enhance both the speed and accuracy of model training. To address these challenges, this study introduces a multi-source remote sensing dataset consisting of 15000 data pairs, each comprising remote sensing multispectral data, synthetic aperture radar(SAR) images, land use and land cover(LULC) data, digital elevation model (DEM) data as well as the analysis data including Slope, Aspect, and Hillshade. Through the application of an end-to-end network based on pix2pix, effective fusion and feature enhancement of multi-source remote sensing data were achieved. The structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and Spectral Angle Mapper(SAM) values reached 0.84, 23.14, and 0.19, respectively, representing significant improvements over the baseline Pix2pix model’s performance metrics of 0.62, 18.84, and 0.23. In downstream semantic segmentation applications, the enhanced dataset was utilized to train semantic segmentation models for remote sensing image analysis. This approach effectively improved the training speed and segmentation accuracy of the models, with the mean intersection over union (mIoU) increasing from 0.467 to 0.481 and accuracy rising from 0.734 to 0.746. Moreover, the visual representation of remote sensing image segmentation demonstrated noticeable enhancements.

Annotated Dataset for Training Cloud Segmentation Neural Networks Using High-Resolution Satellite Remote Sensing Imagery

The integration of satellite data with deep learning has revolutionized various tasks in remote sensing, including classification, object detection, and semantic segmentation. Cloud segmentation in high-resolution satellite imagery is a critical application within this domain, yet progress in developing advanced algorithms for this task has been hindered by the scarcity of specialized datasets and annotation tools. This study addresses this challenge by introducing CloudLabel, a semi-automatic annotation technique leveraging region growing and morphological algorithms including flood fill, connected components, and guided filter. CloudLabel v1.0 streamlines the annotation process for high-resolution satellite images, thereby addressing the limitations of existing annotation platforms which are not specifically adapted to cloud segmentation, and enabling the efficient creation of high-quality cloud segmentation datasets. Notably, we have curated the Annotated Dataset for Training Cloud Segmentation (ADTCS) comprising 32,065 images (512 × 512) for cloud segmentation based on CloudLabel. The ADTCS dataset facilitates algorithmic advancement in cloud segmentation, characterized by uniform cloud coverage distribution and high image entropy (mainly 5–7). These features enable deep learning models to capture comprehensive cloud characteristics, enhancing recognition accuracy and reducing ground object misclassification. This contribution significantly advances remote sensing applications and cloud segmentation algorithms.

Thin Cloud Removal Generative Adversarial Network Based on Sparse Transformer in Remote Sensing Images

Thin clouds in Remote Sensing (RS) imagery can negatively impact subsequent applications. Current Deep Learning (DL) approaches often prioritize information recovery in cloud-covered areas but may not adequately preserve information in cloud-free regions, leading to color distortion, detail loss, and visual artifacts. This study proposes a Sparse Transformer-based Generative Adversarial Network (SpT-GAN) to solve these problems. First, a global enhancement feature extraction module is added to the generator’s top layer to enhance the model’s ability to preserve ground feature information in cloud-free areas. Then, the processed feature map is reconstructed using the sparse transformer-based encoder and decoder with an adaptive threshold filtering mechanism to ensure sparsity. This mechanism enables that the model preserves robust long-range modeling capabilities while disregarding irrelevant details. In addition, inverted residual Fourier transformation blocks are added at each level of the structure to filter redundant information and enhance the quality of the generated cloud-free images. Finally, a composite loss function is created to minimize error in the generated images, resulting in improved resolution and color fidelity. SpT-GAN achieves outstanding results in removing clouds both quantitatively and visually, with Structural Similarity Index (SSIM) values of 98.06% and 92.19% and Peak Signal-to-Noise Ratio (PSNR) values of 36.19 dB and 30.53 dB on the RICE1 and T-Cloud datasets, respectively. On the T-Cloud dataset, especially with more complex cloud components, the superior ability of SpT-GAN to restore ground details is more evident.

Automated Detection of Hillforts in Remote Sensing Imagery With Deep Multimodal Segmentation

ABSTRACTRecent advancements in remote sensing and artificial intelligence can potentially revolutionize the automated detection of archaeological sites. However, the challenging task of interpreting remote sensing imagery combined with the intricate shapes of archaeological sites can hinder the performance of computer vision systems. This work presents a computer vision system trained for efficient hillfort detection in remote sensing imagery. Equipped with an adapted multimodal semantic segmentation model, the system integrates LiDAR‐derived LRM images and aerial orthoimages for feature fusion, generating a binary mask pinpointing detected hillforts. Post‐processing includes margin and area filters to remove edge inferences and smaller anomalies. The resulting inferences are subjected to hard positive and negative mining, where expert archaeologists classify them to populate the training data with new samples for retraining the segmentation model. As the computer vision system is far more likely to encounter background images during its search, the training data are intentionally biased towards negative examples. This approach aims to reduce the number of false positives, typically seen when applying machine learning solutions to remote sensing imagery. Northwest Iberia experiments witnessed a drastic reduction in false positives, from 5678 to 40 after a single hard positive and negative mining iteration, yielding a 99.3% reduction, with a resulting F1 score of 66%. In England experiments, the system achieved a 59% F1 score when fine‐tuned and deployed countrywide. Its scalability to diverse archaeological sites is demonstrated by successfully detecting hillforts and other types of enclosures despite their typical complex and varied shapes. Future work will explore archaeological predictive modelling to identify regions with higher archaeological potential to focus the search, addressing processing time challenges.

Modeling cheatgrass distribution, abundance, and response to climate change as a function of soil microclimate.

Exotic annual grass invasions in water-limited systems cause degradation of native plant and animal communities and increased fire risk. The life history of invasive annual grasses allows for high sensitivity to interannual variability in weather. Current distribution and abundance models derived from remote sensing, however, provide only a coarse understanding of how species respond to weather, making it difficult to anticipate how climate change will affect vulnerability to invasion. Here, we derived germination covariates (rate sums) from mechanistic germination and soil microclimate models to quantify the favorability of soil microclimate for cheatgrass (Bromus tectorum L.) establishment and growth across 30 years at 2662 sites across the sagebrush steppe system in the western United States. Our approach, using four bioclimatic covariates alone, predicted cheatgrass distribution with accuracy comparable to previous models fit using many years of remotely-sensed imagery. Accuracy metrics from our out-of-sample testing dataset indicate that our model predicted distribution well (72% overall accuracy) but explained patterns of abundance poorly (R2 = 0.22). Climatic suitability for cheatgrass presence depended on both spatial (mean) and temporal (annual anomaly) variation of fall and spring rate sums. Sites that on average have warm and wet fall soils and warm and wet spring soils (high rate sums during these periods) were predicted to have a high abundance of cheatgrass. Interannual variation in fall soil conditions had a greater impact on cheatgrass presence and abundance than spring conditions. Our model predicts that climate change has already affected cheatgrass distribution with suitable microclimatic conditions expanding 10%-17% from 1989 to 2019 across all aspects at low- to mid-elevation sites, while high- elevation sites (>2100 m) remain unfavorable for cheatgrass due to cold spring and fall soils.

FBA-DPAttResU-Net: Forest burned area detection using a novel end-to-end dual-path attention residual-based U-Net from post-fire Sentinel-1 and Sentinel-2 images

Forest burned area (FBA) detection using remote sensing (RS) data is critical for timely forest management and recovery attempts after wildfires. This study introduces a dual-path attention residual-based U-Net (DPAttResU-Net), a novel end-to-end deep learning (DL) model tailored for FBA detection using dual-source post-fire Sentinel-1 (S1) and Sentinel-2 (S2) satellite RS imagery. To better distinguish FBAs from other land cover types, DPAttResU-Net incorporates a dual-pathway structure to exploit complementary geometrical/physical and spectral features from S1 and S2, respectively. An integral component in the proposed architecture is the channel-spatial attention residual (CSAttRes) block, which emphasizes salient features through the channel and spatial attention modules, thus improving the burned area feature representation. To compare DPAttResU-Net to state-of-the-art DL models, experiments were conducted on benchmark FBA datasets collected over 12 areas, where ten datasets were used as training data and two datasets were used to test the trained DL models. The experimental results demonstrate the high proficiency of the proposed deep model in meticulously delineating FBAs. In further detail, DPAttResU-Net, with a PFN of 17.96 (%) in the first case and an IoU of 89.31 (%) in the second case, outperformed the existing U-Net-based models. Accordingly, through dual-path integration and attention mechanisms, DPAttResU-Net contributes to accurately identifying FBAs by preserving their geometrical details, making it a promising tool for post-wildfire forest management.

A Texture-Considerate Convolutional Neural Network Approach for Color Consistency in Remote Sensing Imagery

Remote sensing allows us to conduct large-scale scientific studies that require extensive mapping and the amalgamation of numerous images. However, owing to variations in radiation, atmospheric conditions, sensor perspectives, and land cover, significant color discrepancies often arise between different images, necessitating color consistency adjustments for effective image mosaicking and applications. Existing methods for color consistency adjustment in remote sensing images struggle with complex one-to-many nonlinear color-mapping relationships, often resulting in texture distortions. To address these challenges, this study proposes a convolutional neural network-based color consistency method for remote sensing cartography that considers both global and local color mapping and texture mapping constrained by the source domain. This method effectively handles complex color-mapping relationships while minimizing texture distortions in the target image. Comparative experiments on remote sensing images from different times, sensors, and resolutions demonstrated that our method achieved superior color consistency, preserved fine texture details, and provided visually appealing outcomes, assisting in generating large-area data products.

ConvMambaSR: Leveraging State-Space Models and CNNs in a Dual-Branch Architecture for Remote Sensing Imagery Super-Resolution

Deep learning-based super-resolution (SR) techniques play a crucial role in enhancing the spatial resolution of images. However, remote sensing images present substantial challenges due to their diverse features, complex structures, and significant size variations in ground objects. Moreover, recovering lost details from low-resolution remote sensing images with complex and unknown degradations, such as downsampling, noise, and compression, remains a critical issue. To address these challenges, we propose ConvMambaSR, a novel super-resolution framework that integrates state-space models (SSMs) and Convolutional Neural Networks (CNNs). This framework is specifically designed to handle heterogeneous and complex ground features, as well as unknown degradations in remote sensing imagery. ConvMambaSR leverages SSMs to model global dependencies, activating more pixels in the super-resolution task. Concurrently, it employs CNNs to extract local detail features, enhancing the model’s ability to capture image textures and edges. Furthermore, we have developed a global–detail reconstruction module (GDRM) to integrate diverse levels of global and local information efficiently. We rigorously validated the proposed method on two distinct datasets, RSSCN7 and RSSRD-KQ, and benchmarked its performance against state-of-the-art SR models. Experiments show that our method achieves SOTA PSNR values of 26.06 and 24.29 on these datasets, respectively, and is visually superior, effectively addressing a variety of scenarios and significantly outperforming existing methods.

TBNet: A texture and boundary-aware network for small weak object detection in remote-sensing imagery

Object detection is of great importance for remote sensing image interpretation work and has received significant attention. However, small weak object detection has always been a challenge. The main reason is that the critical information of these objects, such as textures and boundaries, is suppressed by the background and cannot effectively express their own characteristics. To address this issue, we introduce a novel texture and boundary-aware network (TBNet) in this paper. Firstly, we propose a texture-aware enhancement module (TAEM) to explore the texture details within the images. TAEM captures pixel correlations to perceive the distribution of texture in the objects. Secondly, a boundary-aware fusion module (BAFM) is introduced to emphasize spatial positions. BAFM can extract the edge information to guide the prediction of small weak objects. Finally, a task-decoupled RCNN (TD-RCNN) is designed to separate classification and regression tasks. TD-RCNN achieves fine-grained detection, avoiding compromises between subtasks. Comprehensive experiments on four public datasets, DIOR NWPU VHR-10, RSOD, and AI-TOD, demonstrate that TBNet achieves state-of-the-art performance compared to competitors. The model is also evaluated on UAVOD-10, which collects numerous small weak objects. TBNet achieves state-of-the-art results while significantly outperforming competitors, proving its ability to detect small weak objects.

Ship-VNet: An Algorithm for Ship Velocity Analysis Based on Optical Remote Sensing Imagery Containing Kelvin Wakes

Extracting ship velocity vectors from optical remote sensing images is a very challenging task, and ship wakes are the only motion features of ships. However, because the sensor’s field of view is not sufficiently bright and the brightness is not uniform, the image contains noise, which makes it difficult to define and extract the wake of the ship. Velocity analysis of the extracted wake makes the whole process complicated and slow. Therefore, considering the above problems, this paper proposes Ship-VNet, an optical remote sensing image ship velocity analysis algorithm based on Kelvin wakes. In this model, the rotating target detection algorithm is used to detect the ship, and then, the classical relationship between the kinematic characteristics of the ship’s Kelvin wake and the velocity of the ship is studied and experimentally analyzed in the frequency domain. In addition, based on optical remote sensing images and corresponding real AIS data, a ship dataset with Kelvin wakes marked with heading velocity was constructed to verify the effectiveness of the proposed method. Compared with the ship velocity analysis method based on the frequency domain, which was also used in the previous research, the experiment demonstrates the superiority of the method in terms of analysis accuracy.

Large-scale integration of remotely sensed and GIS road networks: A full image-vector conflation approach based on optimization and deep learning

Road networks play an important role in the sustainable development of human society. Conventionally, there are two sources of road data acquisition: road extraction from Remote Sensing (RS) imagery and GIS based map production. Each method has its limitations. The RS road extraction methods are primarily raster-based and the extracted roads are not directly usable in GIS due to their fragmented and noisy nature, while vector-based methods cannot utilize rich raster information. Further more, the vector and raster data can have discrepancies for various reasons. Efficient road data production requires an image-vector conflation process that can match and combine raster and vector-based road data automatically.In this study, we propose a full image-vector conflation framework that directly integrates image and vector road data by appropriately transforming extracted roads from imagery and establishing a match relation between these roads and a credible target GIS road dataset. Based on analyzing these match relations, we propose new metrics for measuring the degree of agreement between the raster and vector road data. The proposed framework combines state-of-the-art deep learning methods for image segmentation and optimization-based models for object matching. We prepared a large-scale high-resolution road dataset covering two counties in Kansas, US. Using trained models from one of the two counties, we were able to extract road segments in the other county and match them to the TIGER/Line roads.Our experiments show that conventional performance metrics for road extraction (e.g. IoU) are insufficient for measuring the degree of agreement between image and vector roads as they are pixel-based and are too sensitive to spatial displacement. Instead, the newly defined vector-based agreement metrics are needed for image-vector conflation purposes. Experiments show that, by the vector-based metrics, nearly 90% of GIS road lengths in the study area were extracted and over 90% of extracted roads matched the target GIS roads. The new framework streamlines raster-vector conflation of roads and can potentially expedite relevant geospatial analyses regarding change detection, disaster monitoring and GIS data production, among others.

Refined Land Use Classification for Urban Core Area from Remote Sensing Imagery by the EfficientNetV2 Model

In the context of accelerated urbanization, assessing the quality of the existing built environment plays a crucial role in urban renewal. In the existing research and use of deep learning models, most categories are urban construction areas, forest land, farmland, and other categories. These categories are not conducive to a more accurate analysis of the spatial distribution characteristics of urban green space, parking space, blue space, and square. A small sample of refined land use classification data for urban built-up areas was produced using remote sensing images. The large-scale remote sensing images were classified using deep learning models, with the objective of inferring the fine land category of each tile image. In this study, satellite remote sensing images of four cities, Handan, Shijiazhuang, Xingtai, and Tangshan, were acquired by Google Class 19 RGB three-channel satellite remote sensing images to establish a data set containing fourteen urban land use classifications. The convolutional neural network model EfficientNetV2 is used to train and validate the network framework that performs well on computer vision tasks and enables intelligent image classification of urban remote sensing images. The model classification effect is compared and analyzed through accuracy, precision, recall, and F1-score. The results show that the EfficientNetV2 model has a classification recognition accuracy of 84.56% on the constructed data set. The testing set accuracy increases sequentially after transfer learning. This paper verifies that the proposed research framework has good practicality and that the results of the land use classification are conducive to the fine-grained quantitative analysis of built-up environmental quality.