High Spatial Resolution Images Research Articles

Open Foris Collect Earth: a remote sensing sampling survey of Azerbaijan to support climate change reporting in the land use, land use change, and forestry.

Land use, land use change, and forestry (LULUCF) are critical in climate change mitigation. Producing or collecting activity data for LULUCF is essential in developing national greenhouse gas inventories, national communications, biennial update reports, and nationally determined contributions to meet international commitments under climate change. Collect Earth is a free, publicly accessible software for monitoring dynamics between all land use classes: forestlands, croplands, grasslands, wetlands, settlements, and other lands. Collect Earth supports countries in monitoring the trends in land use and land cover over time by applying a sample-based approach and generating reliable, high-quality, consistent, accurate, transparent, robust, comparable, and complete activity data through augmented visual interpretation for climate change reporting. This article reports forest extent estimates in Azerbaijan, analyzing 7782 0.5-ha sampling units through an augmented visual interpretation of very high spatial and temporal resolution images on the Google Earth platform. The results revealed that in 2016, tree cover existed in 31.9% of total land, equal to 2,751,167ha and 1,301,188ha or 15.1% of the total land, with a 5.4% sampling error covered by forests. The estimate is 15 to 25% higher than the previous estimates, equal to 169,418 to 260,888ha of forest that was never reported in previous studies.

Matrix-Enhanced SIMS: The Influence of Primary Ion Species and Cluster Size on Ion Yield and Ion Yield Enhancement of Lipids.

Time-of-flight secondary ion mass spectrometry is one of the most promising techniques for label-free analysis of biomolecules with nanoscale spatial resolution. However, high-resolution imaging of larger biomolecules such as phospholipids and peptides is often hampered by low yields of molecular ions. Matrix-enhanced SIMS (ME-SIMS), in which an organic matrix is added to the sample, is one promising approach to enhancing the ion yield for biomolecules. Optimizing this approach has, however, been challenging because the processes involved in increasing the ion yield in ME-SIMS are not yet fully understood. In this work, the matrix α-cyano-4-hydroxycinnamic acid (HCCA) has been combined with cluster primary ion analysis to better understand the roles of proton donation and reduced fragmentation on lipid molecule ion yield. A model system consisting of 1:100 mol ratio dipalmitoylphosphatidylcholine (DPPC) in HCCA as well as an HCCA-coated mouse brain cryosection have been studied using a range of Bi and Ar cluster ions. Although the molecular ion yield increased with an increase in cluster ion size, the enhancement of the signals from intact lipid molecules decreased with an increase in cluster ion size for both the model system and the mouse brain. Additionally, in both systems, protonated molecular ions were significantly more enhanced than sodium and potassium cationized molecules for all of the primary ions utilized. For the model system, the DPPC molecular ion yield was increased by more than an order of magnitude for all of the primary ions studied, and fragmentation of DPPC was dramatically reduced. However, on the brain sample, even though the HCCA matrix reduced DPPC fragmentation for all of the primary ions studied, the matrix coating suppressed the ion yield for some lipids when the larger cluster primary ions were employed. This indicated insufficient migration of the lipids into the matrix coating, so that dilution by the matrix overpowered the enhancement effect. This study provides strong evidence that the HCCA matrix both enhances protonation and reduces fragmentation. For imaging applications, the ability of the analytes to migrate to the surface of the matrix coating is also a critical factor for useful signal enhancement. This work demonstrates that the HCCA matrix provides a softer desorption environment when using Bi cluster ions than that obtained using the large gas cluster ions studied alone, indicating the potential for improved high spatial resolution imaging with ME-SIMS.

A Novel Multi-Model High Spatial Resolution Method for Analysis of DCE MRI Data: Insights from Vestibular Schwannoma Responses to Antiangiogenic Therapy in Type II Neurofibromatosis.

This study aimed to develop and evaluate a new DCE-MRI processing technique that combines LEGATOS, a dual-temporal resolution DCE-MRI technique, with multi-kinetic models. This technique enables high spatial resolution interrogation of flow and permeability effects, which is currently challenging to achieve. Twelve patients with neurofibromatosis type II-related vestibular schwannoma (20 tumours) undergoing bevacizumab therapy were imaged at 1.5 T both before and at 90 days following treatment. Using the new technique, whole-brain, high spatial resolution images of the contrast transfer coefficient (Ktrans), vascular fraction (vp), extravascular extracellular fraction (ve), capillary plasma flow (Fp), and the capillary permeability-surface area product (PS) could be obtained, and their predictive value was examined. Of the five microvascular parameters derived using the new method, baseline PS exhibited the strongest correlation with the baseline tumour volume (p = 0.03). Baseline ve showed the strongest correlation with the change in tumour volume, particularly the percentage tumour volume change at 90 days after treatment (p < 0.001), and PS demonstrated a larger reduction at 90 days after treatment (p = 0.0001) when compared to Ktrans or Fp alone. Both the capillary permeability-surface area product (PS) and the extravascular extracellular fraction (ve) significantly differentiated the 'responder' and 'non-responder' tumour groups at 90 days (p < 0.05 and p < 0.001, respectively). These results highlight that this novel DCE-MRI analysis approach can be used to evaluate tumour microvascular changes during treatment and the need for future larger clinical studies investigating its role in predicting antiangiogenic therapy response.

TCNet: A Transformer–CNN Hybrid Network for Marine Aquaculture Mapping from VHSR Images

Precise delineation of marine aquaculture areas is vital for the monitoring and protection of marine resources. However, due to the coexistence of diverse marine aquaculture areas and complex marine environments, it is still difficult to accurately delineate mariculture areas from very high spatial resolution (VHSR) images. To solve such a problem, we built a novel Transformer–CNN hybrid Network, named TCNet, which combined the advantages of CNN for modeling local features and Transformer for capturing long-range dependencies. Specifically, the proposed TCNet first employed a CNN-based encoder to extract high-dimensional feature maps from input images. Then, a hierarchical lightweight Transformer module was proposed to extract the global semantic information. Finally, it employed a coarser-to-finer strategy to progressively recover and refine the classification results. The results demonstrate the effectiveness of TCNet in accurately delineating different types of mariculture areas, with an IoU value of 90.9%. Compared with other state-of-the-art CNN or Transformer-based methods, TCNet showed significant improvement both visually and quantitatively. Our methods make a significant contribution to the development of precision agricultural in coastal regions.

Multifunctional Near-Infrared Dye IR-817 Encapsulated in Albumin Nanoparticles for Enhanced Imaging and Photothermal Therapy in Melanoma.

Near-infrared cyanine dyes have high sensitivity and spatial resolution imaging capabilities, but they also have unavoidable drawbacks such as photobleaching, low water solubility, fluorescence quenching, and toxic side effects. As an effective biologic drug carrier, albumin combines with cyanine dyes to form albumin@dye nanoparticles. These nanoparticles can alleviate the aforementioned issues and are widely used in tumor imaging and photothermal therapy. Herein, a newly synthesized near-infrared dye IR-817 was combined with bovine serum albumin (BSA) to create BSA@IR-817 nanoparticles. Through the detection of fluorescence emission and absorption, the optimal concentration and ratio of BSA and IR-817 were determined. Subsequently, dynamic light scattering (DLS) measurements and scanning electron microscopy (SEM) were used for the physical characterization of the BSA@IR-817 nanoparticles. Finally, in vitro and in vivo experiments were conducted to assess the fluorescence imaging and photothermal therapeutic potential of BSA@IR-817 nanoparticles. IR-817 was adsorbed onto the BSA carrier by covalent conjugation and supramolecular encapsulation, resulting in the formation of dispersed, homogeneous, and stable nanoparticles with a particle size range of 120-220 nm. BSA@IR-817 not only improved the poor water solubility, fluorescence quenching, and toxic side effects of IR-817 but also enhanced the absorption and fluorescence emission peaks in the near-infrared region, as well as the fluorescence in the visible spectrum. In addition, BSA@IR-817 combined with laser 808 irradiation was able to convert light energy into heat energy with temperatures exceeding 50 °C. By creating a mouse model of subcutaneous melanoma, it was discovered that the tumor inhibition rate of BSA@IR-817 was greater than 99% after laser irradiation and that it achieved nearly complete tumor ablation without causing significant toxicity. Our research, therefore, proposes the use of safe and effective photothermal nanoparticles for the imaging, diagnosis, and treatment of melanoma, and offers a promising strategy for future biomedical applications.

Spatial resolution improvement of the multispectral satellite images in the absence of high spatial resolution images through integration of image fusion and deep learning methods

Spatial resolution improvement of the multispectral satellite images in the absence of high spatial resolution images through integration of image fusion and deep learning methods

(Invited) Multiphysics Model of the sSNOM Spectroscopy: Can Near-Field Probes Report the Spectrum of Nanomaterials?

Last decade has been celebrated with a substantial increase of publications on imaging of nanomaterials with near-field probes. The latter include Tip-Enhanced Raman and/or Photoluminescence Spectroscopy (TERS/TEPL), various near-infrared (NIR) Atomic Force Microscopy (AFM) based methods, and several scattering Scanning Near-field optical methods (sSNOM). Using various coherent sources, often tunable in terms of excitation, these methods produce high spatial resolution images – near-field scanning maps, in terms of a pixel size claimed to be limited only by the size of the scanning tip.By design, TERS/TEPL detects a spectral response at each point of a map, which is, naively speaking, gives us a local probe of optical material properties within a single pixel (at the fixed excitation wavelength). NIR-AFM and sSNOM (leaving aside the FTIR mode) allow to obtain quasi-spectral information by tweaking the excitation. In this case, the collected (elastic) signal is restricted to the wavelength of the excitation.The essence of the near-field method is in breaking the laws of a plane-wave reflection from the sample, which requires a sub-wavelength separation between the tip and the sample and a strong enough near-field coupling between those. Still, it is routinely assumed that the tip is an “ideal probe” and its role is to “channel” the information about the local optical response to the detector. Up to date, little is known about the spectral response of the tip-sample system from the perspective of quantum physics.In this work, several aspects of the problem: mechanical motion of the scanning tip, electrodynamics of the excitation/scattered light, and quantum solid state description of the sample and the tip – are combined to obtain, within a simple model, a full picture of the near-field detection. The results, though not general and limited by the simplicity of model used for the sake of clarity of derivation, shed the light on the spectral response of a typical near-field microscope and describe effects, earlier called “artifacts”, as natural response for coupled tip-sample elementary excitations. Lastly, an application of the model for an analysis of spectra of nBN heteronanotubes (Y. Feng, et.al, ACS Nano 15, 5600, 2021) will be presented. Acknowledgement: This work has been partially supported by NSF MRSEC (DMR-2011839).Fig. 1. (left) Amplitude and phase of the near-field optical signal vs. the optical frequency: S(ω). (right) Same optical amplitude and phase vs. ω and the phase of mechanical motion of an AFM tip. Figure 1

Two-Dimensional Population Receptive Field Mapping of Human Primary Somatosensory Cortex.

Functional magnetic resonance imaging can provide detailed maps of how sensory space is mapped in the human brain. Here, we use a novel 16 stimulator setup (a 4 × 4 grid) to measure two-dimensional sensory maps of between and within-digit (D2-D4) space using high spatial-resolution (1.25mm isotropic) imaging at 7 Tesla together with population receptive field (pRF) mapping in 10 participants. Using a 2D Gaussian pRF model, we capture maps of the coverage of digits D2-D5 across Brodmann areas and estimate pRF size and shape. In addition, we compare results to previous studies that used fewer stimulators by constraining pRF models to a 1D Gaussian Between Digit or 1D Gaussian Within Digit model. We show that pRFs across somatosensory areas tend to have a strong preference to cover the within-digit axis. We show an increase in pRF size moving from D2-D5. We quantify pRF shapes in Brodmann area (BA) 3b, 3a, 1, 2 and show differences in pRF size in Brodmann areas 3a-2, with larger estimates for BA2. Generally, the 2D Gaussian pRF model better represents pRF coverage maps generated by our data, which itself is produced from a 2D stimulation grid.

Recent progress in the application of in situ atomic force microscopy for metal anode processes in energy storage batteries

Metal anodes are considered promising candidates for next-generation rechargeable batteries owing to their high theoretical specific capacities. However, practical applications are limited by safety concerns and poor electrochemical performance caused by unstable solid electrolyte interphase (SEI) and uncontrolled metal deposition at the metal anode/electrolyte interface. An in-depth understanding of the interfacial reactions is of vital significance for the development of metal anode-based batteries. In situ electrochemical atomic force microscopy (EC-AFM) enabling high spatial resolution imaging and multifunctional detection is widely used to monitor electrode/electrolyte interfaces in working batteries. In this review, we summarize recent advances in the application of in situ EC-AFM for metal anode processes, including SEI formation and the deposition/dissolution processes of metallic lithium, magnesium, and zinc in metal anode-based batteries, which are conducive to the optimization of metal anodes in energy storage batteries.

Spatial enhanced spatiotemporal reflectance fusion model for heterogeneous regions with land cover change

Numerous spatiotemporal fusion models have been developed to fuse dense time-series data with a high spatial resolution for monitoring land surface dynamics. Nonetheless, enhancing spatial details of fused images, eliminating the obvious ‘plaque’ phenomenon and image blurring in fused images, and developing relatively simple and 1 easy-to-implement algorithms remain a challenge for spatiotemporal fusion algorithms. Therefore, this paper presents a newly proposed spatial enhanced spatiotemporal reflectance fusion model (SE-STRFM) for image fusions in heterogeneous regions with land cover change. The SE-STRFM model predicts temporal changes of reflectance in sub-pixel details based on the spectral unmixing theory, and allocates reflectance changes caused by abrupt land cover change in fine-resolution images with a relatively simple algorithm and easy implementation. SE-STRFM only needs one pair of input data, comprising one fine-resolution image and one coarse-resolution image, to achieve high-precision reflectance prediction with spatial details. To verify the reliability and applicability of the SE-STRFM, we use Landsat image and simulated MODIS-like image to fuse high spatial and temporal resolution images and select two study areas with heterogeneous landscape and land cover type change for fusion experiments and accuracy evaluation. The results show that the images fused by SE-STRFM have clearer spatial details and a more accurate spectral distribution compared with those fused by the most widely used STARFM, ESTARFM and FSDAF. In two study areas with heterogeneous landscape and land cover type change, compared with STARFM, ESTARFM and FSDAF, the RMSE of SE-STRFM is 10.52%, 28.39% and 6.58% lower on average, respectively; r is 3.67%, 10.33% and 1.65% higher on average, respectively; AAD is 9.05%, 24.58% and 7.29% lower on average, respectively; and SSIM is 3.16%, 10.16% and 1.92% higher on average, respectively. SE-STRFM can accurately capture temporal changes with spatial details and effectively predict abrupt land-cover changes.

Optical-scanning evanescent sensor for high-speed broadband photoacoustic microscopy

Photoacoustic microscopy with high spatial resolution and fast imaging acquisition allows observing dynamic processes of optical absorption-based microanatomic structures in three dimensions. An evanescent sensor accesses ultrasonic detection with high sensitivity and broad bandwidth while suffering from limited field of view (FOV), thus compromising the photoacoustic imaging acquisition rate. Here, we develop an optical-scanning evanescent sensor by fast deflection of the interrogation light along the interface of prism and water using a one-dimensional galvanometer, demonstrating excellent detection sensitivity of ∼132 Pa with a broadband frequency response of &amp;gt;140-MHz at an enlarged FOV of ∼2.90 × 0.19 mm2. Incorporating the optical-scanning evanescent sensor in photoacoustic microscopy, a volumetric image (∼3.0 × 0.25 × 1.0 mm3) with micrometer-scale spatial resolution is acquired within ∼2.5 s by synergistically scanning both photoacoustic illumination laser and sensor's interrogation light. High-speed imaging of flowing microparticles within a capillary tube offers the visualizations of the traveling processes in three dimensions. Potentially, the optical-scanning evanescent sensor allows photoacoustic microscopy accommodating to dynamic imaging at cellular level such as in vivo flow cytometry of circulating tumor cells.

Carbon Coated Iron-Cobalt Nanoparticles for Magnetic Particle Imaging.

Magnetic particle imaging (MPI) is an emerging imaging modality that provides direct and quantitative mapping of iron oxide tracers. To achieve high sensitivity and good spatial resolution images, a magnetic nanoparticle with a higher contrast intensity needs to be developed. Currently, a majority of MPIs being developed for potential clinical application are composed of iron oxide nanoparticles with a spherical shape. In this project, we intend to report development of high-performance carbon (C) coated iron-cobalt (FeCo) nanoparticles (FeCo/C) and investigate their feasibility as a MPI agent. We have synthesized FeCo/C through a facile and simple method at mild temperature that is safe, easy, and up-scalable. We studied the structural and functional relationships and biocompatibility of this MPI agent in vitro. However, to enhance the aqueous solubility and biocompatibility, the surface of FeCo/C was modified with polyethylene glycol (PEG). We found that variation in the ratio of Fe and Co plays a vital role in their physical properties and functionality. In vitro imaging confirms that the Fe3Co1/C nanoparticle has highly competitive MPI intensity compared to VivoTrax, a commercially available MPI agent. Confocal laser scanning microscopy imaging with Rhodamine B labeled FeCo/C displays cellular internalization by the A375 cancer cells. The in vitro toxicity analysis concludes that there is no significant toxicity of FeCo/C nanoparticles. Therefore, the newly developed MPI agent holds strong promise for biomedical imaging and could be further validated in vivo in small animals.

Initial Clinical Images From a Second-Generation Prototype Silicon-Based Photon-Counting Computed Tomography System

To demonstrate the feasibility andpotential of using a second-generation prototype photon-counting computed tomography (CT) system to provide simultaneous high spatial resolution images and high spectral resolution material information across a range of routine imaging tasks using clinical patient exposure levels. The photon-counting system employs an innovative silicon-based photon-counting detector to provide a balanced approach to ultra-high-resolution spectral CT imaging. An initial cohort of volunteer subjects was imaged using the prototype photon-counting system. Acquisition technique parameters and radiation dose exposures were guided by routine clinical exposure levels used at the institution. Images were reconstructed in native slice thickness using an early version of a spectral CT reconstruction algorithm Samples of images across a range of clinical tasks were selected and presented for review. Clinical cases are presented across inner ear, carotid angiography, chest, and musculoskeletal imaging tasks. Initial reconstructed images illustrate ultra-high spatial resolution imaging. The fine detail of small structures and pathologies is clearly visualized, and structural boundaries are well delineated. The prototype system additionally provides concomitant spectral information with high spatial resolution. This initial study demonstrates that routine imaging at clinically appropriate patient exposure levels is feasible using a novel deep-silicon photon-counting detector CT system. Furthermore, a deep-silicon detector may provide a balanced approach to photon-counting CT, providing high spatial resolution imaging with simultaneous high-fidelity spectral information.

Non-Paraxial Effects in the Laser Beams Sharply Focused to Skin Revealed by Unidirectional Helmholtz Equation Approximation

Laser beams converging at significant focusing angles have diverse applications, including quartz-enhanced photoacoustic spectroscopy, high spatial resolution imaging, and profilometry. Due to the limited applicability of the paraxial approximation, which is valid solely for smooth focusing scenarios, numerical modeling becomes necessary to achieve optimal parameter optimization for imaging diagnostic systems that utilize converged laser beams. We introduce a novel methodology for the modeling of laser beams sharply focused on the turbid tissue-like scattering medium by employing the unidirectional Helmholtz equation approximation. The suggested modeling approach takes into account the intricate structure of biological tissues, showcasing its ability to effectively simulate a wide variety of random multi-layered media resembling tissue. By applying this methodology to the Gaussian-shaped laser beam with a parabolic wavefront, the prediction reveals the presence of two hotspots near the focus area. The close-to-maximal intensity hotspot area has a longitudinal size of about 3–5 μm and a transversal size of about 1–2 μm. These values are suitable for estimating spatial resolution in tissue imaging when employing sharply focused laser beams. The simulation also predicts a close-to-maximal intensity hotspot area with approximately 1 μm transversal and longitudinal sizes located just behind the focus distance for Bessel-shaped laser beams with a parabolic wavefront. The results of the simulation suggest that optical imaging methods utilizing laser beams with a wavefront produced by an axicon lens would exhibit a limited spatial resolution. The wavelength employed in the modeling studies to evaluate the sizes of the focus spot is selected within a range typical for optical coherence tomography, offering insights into the limitation of spatial resolution. The key advantage of the unidirectional Helmholtz equation approximation approach over the paraxial approximation lies in its capability to simulate the propagation of a laser beam with a non-parabolic wavefront.

Hybrid full-pose parameter calibration of a freeform illuminator for Fourier ptychographic microscopy.

As a typical computational method, Fourier ptychographic microscopy (FPM) can realize high spatial resolution and quantitative phase imaging while preserving the large field of view with a low numerical aperture (NA) objective. A programmable light-emitting diode (LED) array is used as a typical illuminator in an FPM system, and the illumination parameters of each LED element are crucial to the success of the FPM reconstruction algorithm. Compared with LED arrays arranged in rectangular arrays, LED arrays with special structures such as domes or rings can effectively improve FPM imaging results and imaging efficiency. As a trade-off, their calibration difficulty is greatly increased due to the lack of geometric constraints of rectangular arrays. In this paper, we propose an effective hybrid full-pose parameter calibration method for freeform LED array illuminators, combining stereoscopic 3D imaging techniques and the geometric constraints of the microscopic platform. First, a stereovision system is used to obtain the accurate 3D position of each LED element of the freeform illuminator and to construct a rigid 3D coordinate LED array system. Then, calibration between the coordinate system of the LED array and that of the optical imaging component is realized according to the geometric features of the brightfield-to-darkfield edges. Finally, we verify the feasibility and effectiveness of the proposed method through full-pose parameter calibration of LED arrays with different arrangement rules.

A Sub-Bottom Type Adaption-Based Empirical Approach for Coastal Bathymetry Mapping Using Multispectral Satellite Imagery

Accurate bathymetric data in shallow water is of increasing importance for navigation safety, coastal management, and marine transportation. Satellite-derived bathymetry (SDB) is widely accepted as an effective alternative to conventional acoustic measurements in coastal areas, providing high spatial and temporal resolution combined with extensive repetitive coverage. Many previous empirical SDB approaches are unsuitable for precision bathymetry mapping in various scenarios, due to the assumption of homogeneous bottom over the whole region, as well as the neglect of various interfering factors (e.g., turbidity) causing radiation attenuation. Therefore, this study proposes a bottom-type adaption-based SDB approach (BA-SDB). Under the consideration of multiple factors including suspended particulates and phytoplankton, it uses a particle swarm optimization improved LightGBM algorithm (PSO-LightGBM) to derive depth of each pre-segmented bottom type. Based on multispectral images of high spatial resolution and in situ observations of airborne laser bathymetry and multi-beam echo sounder, the proposed approach is applied in shallow water around Yuanzhi Island, and achieves the highest accuracy with an RMSE value of 0.85 m compared to log-ratio, multi-band, and classical machine learning methods. The results of this study show that the introduction of water-environment parameters improves the performance of the machine learning model for bathymetric mapping.

A Comparative Study of the CNN Based Models Used for Remote Sensing Image Classification

Remotely sensed images, their classification and accuracy play a vital role in measuring a country’s scientific growth and technological development. Remote Sensing (RS) can be interpreted as a way of assessing the characteristics of a surface or an entity from a distance. This task of identifying and classifying datasets of RS images can be done using Convolutional Neural Network (CNN). For classifying images of large-scale areas, the traditional CNN approach produces coarse maps. For addressing this issue, Object based CNN method can be used. Classifying images with high spatial resolution can be done effectively using Object based image analysis. Deep learning methods offer the strength of auto learning the spatial features of an image. Object scale based adaptive CNN is a novel technique that can improve the accuracy of image classification of high spatial resolution images. For efficient RS image classification, a novel Deep learning approach called distributed CNN can be used which leads to enhanced accuracy of RS image classification. In this paper, three CNN models have been compared while considering the training time and efficiency to classify RS images as parameters of measure to assess the CNN models.

Evolution of gully erosion and susceptibility factors in the urban watershed of the Kimemi (Butembo/DR Congo)

Gully erosion is one of the most intense landscape degradation mechanisms in areas with varying environmental characteristics. Both natural and anthropogenic factors affect the gullying process. Reliable documentation of these processes in tropical African cities is scarce. This study assessed the gully erosion dynamic and the susceptibility factors in the urban watershed of Kimemi in eastern Democratic Republic of Congo (DRC). Data were obtained through a combination of fieldwork and digitization of very high spatial resolution images from Google Earth (from 2011 to 2021). The length, width, and area of large gullies (width ≥ 5 m) were measured for each year of the study. A logistic regression model (LRM) was also used to investigate the influence of both physical and anthropogenic factors on gully susceptibility. The results revealed that the number of gullies has increased from 36 to 61 during the last decade. The gully mean length of 63.9 ± 61.1 m, 129.3 ± 104.9 m, and 174.7 ± 153.8 m were obtained for the years 2011, 2015, and 2021 respectively. The average density of gully network for the study period was 0.12 km/km², while the degraded land was ∼1.3 and ∼1.1 ha/year for 2011–2015 and 2015–2021 for the entire watershed. The significant changes in morphometric parameters (length, width, area) were found only in the bare land and building land uses. A strong and positive relationship between the length (m) and the area (ha) was found. Furthermore, the susceptibility of gullying was significantly influenced by the slope, stream power index (SPI), distance to roads and rivers, land use and land cover (LULC), and normalized difference vegetation index (NDVI). This means the areas located in the bare land and building or close to roads and/or streams are more likely to be gullied. The findings emphasize the impact of urbanization on gully erosion in the Kimemi watershed, highlighting the importance of informed land management decisions with a close attention to anthropogenic factors.

DSFA: cross-scene domain style and feature adaptation for landslide detection from high spatial resolution images

ABSTRACT Rapid and accurate landslide inventory mapping is significant for emergency rescue and post-disaster reconstruction. Nowadays, deep learning methods exhibit excellent performance in supervised landslide detection. However, due to differences between cross-scene images, the performance of existing methods is significantly degraded when directly applied to another scene, which limits the application of rapid landslide inventory mapping. In this study, we propose a novel Domain Style and Feature Adaptation (DSFA) method for cross-scene landslide detection from high spatial resolution images, which can leverage labeled source domain images and unlabeled target domain images to mine robust landslide representations for different scenes. Specifically, we mitigate the large discrepancy between domains at the dataset level and feature level. At the dataset level, we introduce a domain style adaptation strategy to shift landslide styles, which not only bridges the domain gap, but also increases the diversity of landslide samples. At the feature level, adversarial learning and domain distance minimization are integrated to narrow large feature distribution discrepancies for learning domain-invariant information. In addition, to avoid information omission, we improve the U-Net3+ model. Extensive experimental results demonstrate that DSFA has superior detection capability and outperforms other methods, showing its great application potential in unsupervised landslide domain detection.

Disentangling the Benefits of Self-Supervised Learning to Deployment-Driven Downstream Tasks of Satellite Images (Student Abstract)

In this paper, we investigate the benefits of self-supervised learning (SSL) to downstream tasks of satellite images. Unlike common student academic projects, this work focuses on the advantages of the SSL for deployment-driven tasks which have specific scenarios with low or high-spatial resolution images. Our preliminary experiments demonstrate the robust benefits of the SSL trained by medium-resolution (10m) images to both low-resolution (100m) scene classification case (4.25%↑) and very high-resolution (5cm) aerial image segmentation case (1.96%↑), respectively.