ISPRS Journal of Photogrammetry and Remote Sensing最新文献

As a payload of Tianwen-1 (TW-1), the Mars Mineral Spectrometer (MMS) is tasked with acquiring hyperspectral data of the Martian surface to detect material composition. Microdeformations in optical, mechanical, and thermal components result in the MMS experiencing spectral response distortion in orbit, leading to systematic changes in pixel central wavelengths and full width at half maximum (FWHM). Known as the spectral smile, this distortion compromises the accuracy of reflectance inversion and material composition detection. This study introduces a method for detecting the spectral smile through the Martian atmospheric absorption channel, capitalizing on the distinct characteristics of the atmospheric composition and absorption patterns of Mars. A suitable technical route for in-orbit spectral smile detection was established and tested using simulation experiments and MMS-acquired hyperspectral data. Results suggest that the proposed method can attain central wavelength shifts with a maximum error of 0.32 nm and FWHM variations with a maximum error of 1.95 nm. Employing in-orbit spectral smile detection markedly enhances the correction of Martian atmospheric absorption and provides technical support for Martian surface reflectance inversion. https://github.com/wubingnote/MMS-Spectral-Smile.

Underwater images suffer from blurred details and color distortion due to light attenuation from scattering and absorption. Current underwater image enhancement (UIE) methods overlook the effects of forward scattering, leading to difficulties in addressing low contrast and blurriness. To address the challenges caused by forward and backward scattering, we propose a novel variational-based adaptive method for removing scattering components. Our method addresses both forward and backward scattering and effectively removes interference from suspended particles, significantly enhancing image clarity and contrast for underwater applications. Specifically, our method employs a backward scattering pre-processing method to correct erroneous pixel interferences and histogram equalization to remove color bias, improving image contrast. The backward scattering noise removal method in the variational model uses horizontal and vertical gradients as constraints to remove backward scattering noise. However, it can remove a small portion of forward scattering components caused by light deviation. We develop an adaptive method using the Manhattan Distance to completely remove forward scattering. Our approach integrates prior knowledge to construct penalty terms and uses a fast solver to achieve strong decoupling of incident light and reflectance. We effectively enhance image contrast and color correction by combining variational methods with histogram equalization. Our method outperforms state-of-the-art methods on the UIEB dataset, achieving UCIQE and URanker scores of 0.636 and 2.411, respectively.

Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) plays a pivotal role in civilian and military applications. However, the limited labeled samples present a significant challenge in deep learning-based SAR ATR. Few-shot learning (FSL) offers a potential solution, but models trained with limited samples may produce a high probability of incorrect results that can mislead decision-makers. To address this, we introduce uncertainty estimation into SAR ATR and propose Prior knowledge-guided Evidential Deep Learning (Prior-EDL) to ensure reliable recognition in FSL. Inspired by Bayesian principles, Prior-EDL leverages prior knowledge for improved predictions and uncertainty estimation. We use a deep learning model pre-trained on simulated SAR data to discover category correlations and represent them as label distributions. This knowledge is then embedded into the target model via a Prior-EDL loss function, which selectively uses the prior knowledge of samples due to the distribution shift between simulated data and real data. To unify the discovery and embedding of prior knowledge, we propose a framework based on the teacher-student network. Our approach enhances the model’s evidence assignment, improving its uncertainty estimation performance and target recognition accuracy. Extensive experiments on the MSTAR dataset demonstrate the effectiveness of Prior-EDL, achieving recognition accuracies of 70.19% and 92.97% in 4-way 1-shot and 4-way 20-shot scenarios, respectively. For Out-Of-Distribution data, Prior-EDL outperforms other uncertainty estimation methods. The code is available at https://github.com/Xiaoyan-Zhou/Prior-EDL/.

Accurate quantification of snow cover facilitates the prediction of snowmelt runoff, the assessment of freshwater availability, and the analysis of Earth’s energy balance. Existing fractional snow cover (FSC) data, however, often suffer from limitations such as spatial and temporal gaps, compromised accuracy, and coarse spatial resolution. These limitations significantly hinder the ability to monitor snow cover dynamics effectively. To address these formidable challenges, this study introduces a novel data fusion framework specifically designed to generate high-resolution (1 km) daily FSC estimation across vast regions like North America, regardless of weather conditions. It achieved this by effectively integrating the complementary spatiotemporal characteristics of both coarse- and fine-resolution FSC data through a multi-stage processing pipeline. This pipeline incorporates innovative strategies for bias correction, gap filling, and consideration of dynamic characteristics of snow cover, ultimately leading to high accuracy and high spatiotemporal completeness in the fused FSC data. The accuracy of the fused FSC data was thoroughly evaluated over the study period (September 2015 to May 2016), demonstrating excellent consistency with independent datasets, including Landsat-derived FSC (total 24 scenes; RMSE=6.8–18.9 %) and ground-based snow observations (14,350 stations). Notably, the fused data outperforms the widely used Interactive Multi-sensor Snow and Ice Mapping System (IMS) daily snow cover extent data in overall accuracy (0.92 vs. 0.91), F1_score (0.86 vs. 0.83), and Kappa coefficient (0.80 vs. 0.77). Furthermore, the fused FSC data exhibits superior performance in accurately capturing the intricate daily snow cover dynamics compared to IMS data, as confirmed by superior agreement with ground-based observations in four snow-cover phenology metrics. In conclusion, the proposed data fusion framework offers a significant advancement in snow cover monitoring by generating high-accuracy, spatiotemporally complete daily FSC maps that effectively capture the spatial and temporal variability of snow cover. These FSC datasets hold substantial value for climate projections, hydrological studies, and water management at both global and regional scales.

Novel view synthesis (NVS) of urban scenes enables the exploration of cities virtually and interactively, which can further be used for urban planning, navigation, digital tourism, etc. However, many current NVS methods require a large amount of images from known views as input and are sensitive to intrinsic and extrinsic camera parameters. In this paper, we propose a new unified framework for NVS of urban scenes with fewer required views via the integration of scene priors and the joint optimization of camera parameters under an geometric constraint along with NeRF weights. The integration of scene priors makes full use of the priors from the neighbor reference views to reduce the number of required known views. The joint optimization can correct the errors in camera parameters, which are usually derived from algorithms like Structure-from-Motion (SfM), and then further improves the quality of the generated novel views. Experiments show that our method achieves about $25.375$ dB and $25.512$ dB in average in terms of peak signal-to-noise (PSNR) on synthetic and real data, respectively. It outperforms popular state-of-the-art methods (i.e., BungeeNeRF and MegaNeRF ) by about 2–$4$ dB in PSNR. Notably, our method achieves better or competitive results than the baseline method with only one third of the known view images required for the baseline. The code and dataset are available at https://github.com/Dongber/PriNeRF.

Landslides and flash floods are geomorphic hazards (GH) that often co-occur and interact and frequently lead to societal and environmental impact. The compilation of detailed multi-temporal inventories of GH events over a variety of contrasting natural as well as human-influenced landscapes is essential to understanding their behavior in both space and time and allows to unravel the human drivers from the natural baselines. Yet, creating multi-temporal inventories of these GH events remains difficult and costly in terms of human labor, especially when relatively large regions are investigated. Methods to derive GH location from satellite optical imagery have been continuously developed and have shown a clear shift in recent years from conventional methodologies like thresholding and regression to machine learning (ML) methodologies given their improved predictive performance. However, these current generation ML methodologies generally rely on accurate information on either the GH location (training samples) or the GH timing (pre- and post-event imagery), making them unfit in unexplored regions without a priori information on GH occurrences. Currently, a detection methodology to create multi-temporal GH event inventories applicable in relatively large unexplored areas containing a variety of landscapes does not yet exist. We present a new semi-supervised methodology that allows for the detection of both location and timing of GH event occurrence with optical time series, while minimizing manual user interventions. We use the peak of the cumulative difference to the mean for a multitude of spectral indices derived from open-access, high spatial resolution (10–20 m) Copernicus Sentinel-2 time series and generate a map per Sentinel-2 tile that identifies impacted pixels and their related timing. These maps are used to identify GH event impacted zones. We use the generated maps, the identified GH events impacted zones and the automatically derived timing and use them as training sample in a Random Forest classifier to improve the spatial detection accuracy within the impacted zone. We showcase the methodology on six Sentinel-2 tiles in the tropical East African Rift where we detect 29 GH events between 2016 and 2021. We use 12 of these GH events (totalizing ∼3900 GH features) with varying time of occurrence, contrasting landscape conditions and different landslide to flash flood ratios to validate the detection methodology. The average identified timing of the GH events lies within two to four weeks of their actual occurrence. The sensitivity of the methodology is mainly influenced by the differences in landscapes, the amount of cloud cover and the size of the GH events. Our methodology is applicable in various landscapes, can be run in a systematic mode, and is dependent only on a few parameters. The methodology is adapted for massive computation.

Phase linking technique has shown the ability to mitigate the decorrelation effect on the time series interferometric synthetic aperture radar (InSAR) data. By imposing the temporal phase-closure constraint, this technique reconstructs a consistent phase series from the complex sample coherence matrix (SCM). However, the bias of coherence estimates degrades the performance of phase linking, especially in near-zero coherence environments with limited spatial sample support. In this study, we present a methodology to enhance phase linking, with an emphasis on SCM refinement. The incentive behind this is to shrink the tapered SCM towards a scaled identity matrix by exploiting the inner correlation and coherence loss trend in SCM. This allows debiasing the SCM magnitude even in the presence of small sample size. We demonstrate the performance of this method by simulations and real case studies using Sentinel-1 data over Hawaii island. Results from comprehensive comparisons validate the effectiveness of coherence matrix estimation and the enhancement to phase linking in different coherence scenarios. The source code and sample dataset are available at https://www.mathworks.com/matlabcentral/fileexchange/169553-insar-phase-linking-enhancement-by-scm-refinement.

We present a learning-based approach to reconstructing buildings as 3D polygonal meshes from airborne LiDAR point clouds. What makes 3D building reconstruction from airborne LiDAR difficult is the large diversity of building designs, especially roof shapes, the low and varying point density across the scene, and the often incomplete coverage of building facades due to occlusions by vegetation or the sensor’s viewing angle. To cope with the diversity of shapes and inhomogeneous and incomplete object coverage, we introduce a generative model that directly predicts 3D polygonal meshes from input point clouds. Our autoregressive model, called Point2Building, iteratively builds up the mesh by generating sequences of vertices and faces. This approach enables our model to adapt flexibly to diverse geometries and building structures. Unlike many existing methods that rely heavily on pre-processing steps like exhaustive plane detection, our model learns directly from the point cloud data, thereby reducing error propagation and increasing the fidelity of the reconstruction. We experimentally validate our method on a collection of airborne LiDAR data from Zurich, Berlin, and Tallinn. Our method shows good generalization to diverse urban styles.

Recent advances in mobile mapping systems have greatly enhanced the efficiency and convenience of acquiring urban 3D data. These systems utilize LiDAR sensors mounted on vehicles to capture vast cityscapes. However, a significant challenge arises due to occlusions caused by roadside parked vehicles, leading to the loss of scene information, particularly on the roads, sidewalks, curbs, and the lower sections of buildings. In this study, we present a novel approach that leverages deep neural networks to learn a model capable of filling gaps in urban scenes that are obscured by vehicle occlusion. We have developed an innovative technique where we place virtual vehicle models along road boundaries in the gap-free scene and utilize a ray-casting algorithm to create a new scene with occluded gaps. This allows us to generate diverse and realistic urban point cloud scenes with and without vehicle occlusion, surpassing the limitations of real-world training data collection and annotation. Furthermore, we introduce the Scene Gap Completion Network (SGC-Net), an end-to-end model that can generate well-defined shape boundaries and smooth surfaces within occluded gaps. The experiment results reveal that 97.66% of the filled points fall within a range of 5 centimeters relative to the high-density ground truth point cloud scene. These findings underscore the efficacy of our proposed model in gap completion and reconstructing urban scenes affected by vehicle occlusions.

Distributed Scatterers Interferometry (DS-InSAR) has been widely applied to increase the number of measurement points (MP) in complex mountainous areas with dense vegetation and complicated topography. However, DS-InSAR method adopts batch processing mode. When new observation data acquired, the entire archived data is reprocessed, completely ignoring the existing results, and not suitable for high-performance processing of operational observation data. The current research focuses on the automation of SAR data acquisition and processing optimization, but the core time series analysis method remains unchanged. In this paper, based on the traditional Sequential Estimator proposed by Ansari in 2017, a Recursive Sequential Estimator with Flexible Batches (RSEFB) is improved to divide the large dataset flexibly without requirements on the number of images in each subset. This method updates and processes the newly acquired SAR data in near real-time, and obtains long-time sequence results without reprocessing the entire data archived, helpful to the early warning of landslide disaster in the future. 132 Sentinel-1 SAR images and 44 TerraSAR-X SAR images were utilized to inverse the line of sight (LOS) surface deformation of Xishancun landslide and Huangnibazi landslide in Li County, Sichuan Province, China. RSEFB method is applied to retrieve time-series displacements from Sentinel-1 and TerraSAR-X datasets, respectively. The comparison with the traditional Sequential Estimator and validation through Global Position System (GPS) monitoring data proved the effectiveness and reliability of the RSEFB method. The research shows that Xishancun landslide is in a state of slow and uneven deformation, and the non-sliding part of Huangnibazi landslide has obvious deformation signal, so continuous monitoring is needed to prevent and mitigate possible catastrophic slope failure events.