Jul 2026· The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences· Vol XLIX-B3-2026, pp. 205-212· 1 citation· 14 references
TL;DR
Experimental results demonstrate that the adapted SAM2 model achieves stable segmentation under moderate environmental variability, while degrading under severe visibility loss, consistent across model scales and input resolutions.
Abstract
Abstract. Ground based time lapse cameras provide continuous, high frequency observations of glacier dynamics; however, automated analysis of these image streams remains challenging due to fog, snowfall, lens contamination, and variable illumination. This study investigates the potential of adapting the foundation segmentation model Segment Anything Model 2 (SAM2) for glacier segmentation from ground-based monitoring. To enable integration into automated pipelines, SAM2 is configured in image mode with a learned prompt generation strategy, while fine-tuning is restricted to the prompt encoder and mask decoder. In addition, the internal Intersection over Union (IoU) prediction head is utilized as a confidence estimator to assess segmentation reliability. Experimental results demonstrate that the adapted model achieves stable segmentation under moderate environmental variability, while degrading under severe visibility loss. This stability is consistent across model scales and input resolutions. The confidence estimation further provides a meaningful signal for identifying uncertain predictions, supporting reliability-aware processing in downstream workflows.
Abstract. To robustly support glacier calving monitoring at high temporal resolution and enable future AI-based calving forecasts, this study presents an optimized Multi-Epoch Multi-Imagery (MEMI) strategy for automated 4D point cloud model generation. To date, the dataset comprises over 160,000 images acquired since December 2024 by an autonomous multi-camera system operating at 30 min intervals at Glacier Perito Moreno (GPM), Argentina. Despite high scene variability and harsh environmental conditions, the proposed MEMI workflow effectively addresses constraints imposed by continuous glacier motion and image degradation. The enhanced strategy aims to generate precise dense clouds with high alignment accuracy and computational efficiency, forming the basis for subsequent analysis of glacier front evolution. To achieve this, various parameter configurations are evaluated, including AI-based image masking and adaptive, optimized alignment-adjustment settings. Results from a representative eight-day subset show that variations in the tie point computation strategy lead to measurable differences in alignment-adjustment efficiency, with the best configuration being about 11 % faster than the least efficient one. By contrast, adaptive alignment-adjustment consistently improves alignment accuracy. Moreover, masking enhances both image quality checking and reconstruction quality, and, albeit modestly, improves pre-failure deformation analysis. Furthermore, daily seasonal responses to alignment are observed, as accuracy varies with solar illumination relative to the camera positions. Applying the optimal configuration to 260 MEMI projects in under 42 h produced 518 high-precision dense clouds and detected calving retreat magnitudes of up to 17.5m, demonstrating the robustness and scalability of the proposed MEMI strategy for high-temporal-resolution 4D point cloud generation.
Laura Camila Duran Vergara, Xabier Blanch Górriz, Bindusara Nagathihalli Lokesh et al.· ISPRS Annals of the Photogra...· 1 citation
Glacier segmentation in optical remote sensing imagery remains challenging in complex mountain environments due to fragmented glacier structures, blurred boundaries, seasonal snow confusion, terrain shadows, bare rock, and cloud interference. To address these issues, this study proposes a state-space-guided U-Net framework, termed SFG-UNet, for glacier segmentation in optical remote sensing imagery. The model introduces a long-range state space block in the encoder to enhance global contextual representation, an SSM-guided frequency decoupling and boundary calibration module in the skip pathway to refine low- and high-frequency features, and a semantic-guided full-scale gated fusion module in the decoder to improve selective multi-scale feature aggregation. Experiments were conducted on a self-built Landsat-8/9 glacier dataset from the Animaqing Snow Mountain region and an independent public DL4GAM Alps dataset for external validation. On the Animaqing dataset, SFG-UNet achieved 95.71% accuracy, 95.04% dice, 90.08% kappa, and 90.61% MIoU, outperforming representative CNN-based, attention-based, Transformer-based, frequency-domain, glacier-oriented, and SSM-based segmentation methods. On the external DL4GAM Alps dataset, SFG-UNet also achieved the best overall performance, with 89.86% accuracy, 89.12% dice, 81.02% kappa, and 83.64% MIoU. Seasonal, scenario-based, complexity, ablation, and residual error analyses further demonstrate that SFG-UNet improves glacier continuity, boundary recovery, and robustness under complex optical imaging conditions while maintaining an acceptable computational cost.
Yunzhong Shen, Xiuzai Zhang, Changjun Yang et al.· Engineering Research Express· 0 citations
The first systematic zero-shot evaluation of SAM 2 for aerial building segmentation is presented, establishing SAM 2 as a viable tool for rapid building mapping while highlighting where domain adaptation remains necessary.
Bingning Xiong, Mingyu Ou· Journal of image processing...· 0 citations
The automated detection and classification of surface degradation on historical buildings represents a critical challenge in architectural heritage conservation. Conventional approaches relying on manual inspection or supervised machine learning require extensive annotated datasets and expert involvement, limiting their scalability. This paper presents a novel zero-shot pipeline for degradation segmentation on historical civil architecture, combining UAV-acquired photogrammetric data processed in Agisoft Metashape with Gemma 4 31B, Google DeepMind's flagship open-weight vision language model, running locally via LM Studio, and the Segment Anything Model 2 (SAM2) for pixel-accurate mask generation. The system operates entirely without task-specific training data, producing segmentation masks overlaid on the RGB orthomosaic for expert visual evaluation. A case study on a degraded historical building in Calabria, southern Italy, demonstrates the pipeline's ability to detect and categorize detachment, cracking, and lacunae in a unified, reproducible workflow. Results are evaluated through structured expert visual assessment. The approach offers a replicable, low-cost alternative to supervised segmentation, particularly suited to contexts where labeled data is unavailable.
F. Demarco, Federico De Francesca, Pierpaolo Antonio Fusaro et al.· 2026 IEEE International Work...· 0 citations
Although deep learning-based object detectors have achieved remarkable performance on standard benchmarks, their reliability deteriorates significantly under adverse weather conditions (e.g., rain, snow, and fog) due to severe image degradation and the loss of critical structural information. To address this challenge, this paper proposes GPE-YOLO, a robust detection framework built upon the YOLOv11 architecture that explicitly integrates multiscale edge priors to enhance feature resilience. Specifically, we introduce three novel components: 1) Dual-Stream Gradient-Pooling Stem (DSGPS) serves as an efficient preprocessing unit, utilizing parallel 3D Sobel and spatial convolutions to capture comprehensive low-level features and mitigate noise propagation at the input stage. 2) The Pyramidal Edge Prior Generation Module (PEPGM) extracts multi-scale edge maps from shallow layers and injects them into the backbone network via max pooling, effectively suppressing background interference while preserving structural details. 3) The Dynamic Edge-Aware Feature Pyramid (DE-AFP) incorporates DySample and CSPStage to optimize feature upsampling and downsampling processes, thereby ensuring efficient semantic information transmission and robust multi-scale fusion across the network. Extensive experiments on the VOC2012-CWC dataset demonstrate the superiority of our method. GPE-YOLO achieves an mAP@0.5 of 32.84% and mAP@0.5:0.95 of 24.59%, outperforming the baseline YOLOv11 by substantial margins of 4.85% and 3.86%, respectively. These results validate the potential of GPE-YOLO for reliable deployment in real-world adverse weather scenarios.
Xiaojie Chen, Yifei Zhou, Yiming Zhou et al.· International Conference on...· 0 citations
Unmanned aerial vehicle (UAV) target segmentation remains challenging due to the small size of objects, appearance variations, cluttered backgrounds, and the scarcity of densely annotated data. These factors hinder the performance and practical deployment of lightweight segmentation models in real-world UAV applications. To address this problem, this paper investigates the use of SAM3 (Segment Anything Model 3) as a pseudo-label generator for training compact segmentation networks. Specifically, two supervision paradigms are explored: (i) direct pseudo-supervision using unaltered SAM3-generated masks, and (ii) a refinement strategy that re-applies SAM3 to localized image patches for improved mask quality. Based on these paradigms, a two-stage SAM3-guided pseudo-label generation framework is proposed. In the first stage, SAM3 generates coarse masks for initial object localization. The localized regions are subsequently cropped into patches and processed by SAM3 again to generate fine masks with accurate object boundaries and discard false positives. The resulting coarse and fine masks are then used as pseudo-labels to optimize a lightweight network, termed IPS-Seg, which consists of three components: an IdentityFormer backbone for feature extraction, an Atrous Spatial Pyramid Pooling module for multi-scale context aggregation, and a PixelShuffle-based decoder for spatial resolution recovery. Extensive experiments under multiple supervision settings demonstrate the effectiveness of the proposed framework. The results show that IPS-Seg achieves a favorable trade-off between segmentation accuracy and computational efficiency while benefiting consistently from the proposed pseudo-label generation strategy. These findings highlight the potential of large-scale foundation models as annotation sources for training compact task-specific segmentation networks in low-label vision domains.