Conference Agenda

Cultivated land conversion to built-up area is a core form of farmland non-agriculturalization and a major threat to farmland protection in China. Current remote sensing methods for detecting such changes face two limitations: insufficient integration of domain prior knowledge and the inability of purely data-driven models to achieve both high Precision and Recall. To address these issues, this study proposes a knowledge graph-enhanced change detection method. A multi-scale knowledge analysis framework incorporating feature, scene, and business knowledge layers is constructed to systematically integrate multi-source geographic information into structured semantic representations. A knowledge fusion residual network, KG-MS-ResNet, is designed based on ResNet-18 with modifications to the first convolutional layer for bi-temporal image inputs. TransE embeds geographic indicator knowledge into multi-scale semantic vectors, while a semantic–feature dual-path fusion strategy and a knowledge-guided attention mechanism enable deep coupling between image features and domain knowledge. Experiments in Pei County, Jiangsu Province, show that the proposed method outperforms baseline ResNet across all metrics, with Recall increasing by 4.84 percentage points and F1-score by 0.0752. The results demonstrate that integrating domain knowledge graphs with deep learning significantly improves detection performance, offering a semantically interpretable solution for monitoring cultivated land non-agriculturalization and advancing the integration of knowledge-driven and data-driven approaches in intelligent remote sensing interpretation.

Road change detection is essential for maintaining up-to-date digital maps; however, conventional update processes rely heavily on the manual interpretation of aerial imagery, leading to high labor costs and inconsistent outcomes. To address these limitations, this study proposes an automated road change detection method that integrates aerial orthophoto-based segmentation with geometric boundary deviation analysis. Road areas are first extracted from high-resolution aerial orthophotos using SegFormer, a Transformer based semantic segmentation model. The segmentation results are then converted into vector polygons for geometric analysis. Structural changes, such as newly constructed or removed roads, are detected through a difference-based comparison with historical digital maps. Simultaneously, shape changes are quantitatively analyzed by measuring geometric deviations between road boundaries.

Specifically, vertex-wise distances between corresponding boundaries are computed, and the overall deformation is evaluated using Root Mean Square Error (RMSE), incorporating Z-score-based outlier removal to ensure robustness against noise. Experimental results demonstrate that the proposed method effectively detects both structural changes and subtle geometric variations, including road expansions and boundary shifts. Furthermore, the method enables clear object-level classification of change types, providing a practical and efficient framework for digital map updating workflows.

Accurate building change detection depends on how well building boundaries are delineated, as distortions and merging errors hinder reliable correspondence. In dense urban areas, deep learning models frequently merge adjacent buildings—especially within narrow gaps—producing structural inconsistencies that lead to change detection errors. We propose a post-processing method integrating Local Rank-Based Prior Calibration, which reinterprets Softmax probabilities as percentile-based local ranks, with Graph-Cut refinement for structural correction. The refined mask is matched with historical building data to classify four change types. Experiments using aerial imagery from Seoul show that the method reduces structural errors, lowering under-segmentation from 51.64% to 22.02% and improving IoU from 0.748 to 0.759. In change detection, it increases the mean F1-score from 0.522 to 0.608 and improves all classes, including new construction, whose F1-score rises from 0.269 to 0.707. Ablation studies confirm that calibration and graph-based refinement both contribute to the improvements. These results show that stabilizing segmentation outputs enhances the reliability of building-level change detection in dense urban environments.

OpenStreetMap (OSM) is a representative open geospatial platform that provides free access to major spatial objects, including buildings worldwide, constructed through crowdsourcing-based manual digitization. However, subjective differences among contributors and the absence of unified quality control standards have led to the accumulation of positional offsets and boundary shape errors in building polygons. To address this issue, studies using deep learning-based semantic segmentation for OSM quality improvement have been conducted. Nevertheless, Transformer-based segmentation models exhibit an under-segmentation tendency that merges adjacent buildings into a single object, along with limitations in precise boundary delineation. To overcome these challenges, this study proposes a two-stage framework that integrates SegFormer, which excels in global context recognition, with SAM 2, which is capable of precise boundary segmentation. In the first stage, SegFormer semantically segments building regions from a true orthoimage, and in the second stage, SAM 2 infers object-level precise boundaries using the bounding boxes of OSM polygons as box prompts. The two results are combined into a prior probability map, enabling uncertain boundary regions to be re-evaluated in an unsupervised manner. In experiments conducted over the Suseo-dong area in Gangnam-gu, Seoul, the proposed method achieved a BIoU of 70.40%, an improvement of 23.85 percentage points over OSM building data, with consistent performance gains across all evaluation metrics. This framework offers scalability applicable to any region worldwide without additional label construction, provided that high-resolution true orthoimagery and OSM data are available.

Accurate building footprint extraction is critical for applications ranging from population estimation to disaster management. Although optical imagery provides detailed spectral information, it often struggles with shadows, occlusions, and background clutter in dense urban environments. Lidar data, by contrast, offer precise elevation and structural attributes but face challenges such as variable point density and noise. This study integrates multispectral imagery from the U.S. Department of Agriculture (USDA) National Agriculture Imagery Program (NAIP) with lidar-derived feature height and intensity from the U.S. Geological Survey (USGS) 3D Elevation Program (3DEP) to improve footprint extraction using a U-Net–based deep learning model. A six-band input stack (RGB, near-infrared, height, intensity) was developed, normalized, and tiled for training and evaluation against Microsoft Global Building Footprints (GBF). Results from the Houston, TX test site show that the six-band model achieved a precision of 0.86, recall of 0.88, F1 score of 0.87, and Intersection-over-Union (IoU) of 0.76, consistently outperforming four-band baselines by reducing false positives while maintaining sensitivity. Predictions on withheld Houston tiles confirmed strong within-region generalization, yielded a precision of 0.78, recall of 0.81, F1 score of 0.79, and IoU of 0.66. Qualitative analysis further revealed limitations stemming from both training label quality and vegetation–building confusion. These findings demonstrate the complementary value of integrating spectral and structural information for robust building footprint extraction and how domain adaptation strategies can be used to enhance cross-regional transferability.

This contribution presents a benchmarking study of a lightweight deep learning model for automatic pothole detection in asphalt pavements. Accurate and cost-effective identification of surface distresses is essential for road safety and for prioritising maintenance, especially in cities where traditional visual surveys are still predominant. We adapt and train a compact YOLO-based object detection architecture on a dataset of annotated street-level images, covering different lighting conditions, pavement textures and distress severities. The study evaluates how input resolution, confidence thresholds and data augmentation strategies affect detection performance and inference speed, and compares the lightweight model with heavier state-of-the-art detectors. Results indicate that it is possible to obtain competitive accuracy while maintaining real-time processing capabilities on modest hardware, which is crucial for deployment in mobile inspection platforms such as smartphones, dashcams or low-cost onboard units. The paper discusses opportunities and limitations of integrating deep learning into pavement management systems and outlines perspectives for extending the approach to other types of defects and to larger road networks.