Deep Learning-Based Building Footprint Extraction With Missing Annotations

Kang, Jian; Fernández-Beltrán, Rubén; Sun, Xian; Ni, Jingen

doi:10.1109/lgrs.2021.3072589

Cited by 11 publications

(8 citation statements)

References 22 publications

(23 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then, a three-term joint loss function was proposed: 1) logit adjusted cross-entropy, 2) weighted dice loss, and 3) boundary alignment loss. The obtained results indicate that the proposed loss function preserves the fine-grained structure of building boundaries, effectively discriminates between building and background pixels, and increases F1-scores [50].…”

Section: Segmentation Of Building Pointsmentioning

confidence: 86%

Building Footprint Extraction in Dense Area from LiDAR Data using Mask R-CNN

Mohamed¹,

Mahmoud²,

Moustafa³

et al. 2022

IJACSA

View full text Add to dashboard Cite

Building footprint extraction is an essential process for various geospatial applications. The city management is entrusted with eliminating slums, which are increasing in rural areas. Compared with more traditional methods, several recent research investigations have revealed that creating footprints in dense areas is challenging and has a limited supply. Deep learning algorithms provide a significant improvement in the accuracy of the automated building footprint extraction using remote sensing data. The mask R-CNN object detection framework used to effectively extract building in dense areas sometimes fails to provide an adequate building boundary result due to urban edge intersections and unstructured buildings. Thus, we introduced a modified workflow to train ensemble of the mask R-CNN using two backbones ResNet (34, 101). Furthermore, the results were stacked to fine-grain the structure of building boundaries. The proposed workflow includes data preprocessing and deep learning, for instance, segmentation was introduced and applied to a light detecting and ranging (LiDAR) point cloud in a dense rural area. The outperformance of the proposed method produced better-regularized polygons that obtained results with an overall accuracy of 94.63%.

show abstract

Section: Segmentation Of Building Pointsmentioning

confidence: 86%

Building Footprint Extraction in Dense Area from LiDAR Data using Mask R-CNN

Mohamed¹,

Mahmoud²,

Moustafa³

et al. 2022

IJACSA

View full text Add to dashboard Cite

show abstract

“…Publications in urban studies also cover 6% in land survey management [35] and 25% in urban classification and detection [36][37][38][39][40] particularly in building applications and for Urban Land-Use Classification [41].…”

Section: Land Cover Studiesmentioning

confidence: 99%

“…The authors [140] proposed the effectiveness of semi-supervised adversarial learning methods for handling limited and unannotated high-resolution satellite images. Furthermore, the investigations by [39,156,165,166] sought remedies for challenges stemming from the scarcity of well-annotated pixel-level data and as well as other studies proposed steps on how to tackle instances of class imbalance [65,161,167].…”

mentioning

confidence: 99%

A Review and Meta-analysis of Semantic Segmentation Models in Land Use/ Land Cover Mapping

Ajibola,

Cabral

2024

Preprint

View full text Add to dashboard Cite

Recent advancements in deep learning have spurred the development of numerous novel semantic segmentation models for land cover mapping, showcasing exceptional performance in delineating precise boundaries and producing highly accurate land cover maps. However, to date, no systematic literature review has comprehensively examined semantic segmentation models in the context of land cover mapping. This paper addresses this gap by synthesizing recent advancements in semantic segmentation models for land cover mapping from 2017 to 2023, drawing insights on trends, data sources, model structures, and performance metrics based on a review of 106 extracted articles. Our analysis identifies top journals in the field, including MDPI Remote Sensing, IEEE Journal of Selected Topics in Earth Science, and IEEE Transactions on Geoscience and Remote Sensing, IEEE Geoscience and Remote Sensing Letters, ISPRS Journal Of Photogrammetry And Remote Sensing as the leading journals. We find that research predominantly focuses on land cover, urban areas, precision agriculture, environment, coastal areas, and forests. Geographically, 35.29% of the study areas are located in China, followed by USA (11.76%), France (5.88 %), Spain (4%) and others. Sentinel-2, Sentinel-1, and Landsat satellites emerge as the most commonly used data sources. Benchmark datasets such as ISPRS Vaihingen &amp; Potsdam, LandCover.ai, DeepGlobe, and GID datasets are frequently employed. Model architectures predominantly utilize encoder-decoder, and hybrid convolutional neural network-based structures because of their impressive performances, with limited adoption of transformer-based architectures due to its computational complexity issue, and slow convergence speed. Lastly, this paper highlights existing key research gaps in the field to guide future research directions.

show abstract

“…Our paper experimented without any assumption about the distribution of building ratios and error types. However, if there is prior knowledge such that omission error dominates the noisy map, using a loss function to deal with class imbalance such as weighted loss [67] and focal loss [68] or oversampling a minority class [69] might be an additional remedy to improve performance. Lastly, the IoU values of SFL-U-Net and U-Net were not significantly different in the strong supervision case, as shown in Table I.…”

Section: ) Sfl With Different Types Of Noisementioning

confidence: 99%

Self-Filtered Learning for Semantic Segmentation of Buildings in Remote Sensing Imagery With Noisy Labels

Song

Yang

Jung

2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Not all building labels for training improve the performance of the deep learning model. Some labels can be falsely labeled or too ambiguous to represent their groundtruths, resulting in poor performance of the model. For example, building labels in OpenStreetMap (OSM) and Microsoft Building Footprints (MBF) are publicly available training sources that have great potential to train deep models, but directly using those labels for training can limit the model's performance as their labels are incomplete and inaccurate, called noisy labels. This paper presents self-filtered learning (SFL) that helps a deep model learn well with noisy labels for building extraction in remote sensing images. SFL iteratively filters out noisy labels during the training process based on loss of samples. Through a multi-round manner, SFL makes a deep model learn progressively more on refined samples from which the noisy labels have been removed. Extensive experiments with the simulated noisy map as well as real-world noisy maps, OSM and MBF, showed that SFL can improve the deep model's performance in diverse error types and different noise levels.

show abstract

Deep Learning-Based Building Footprint Extraction With Missing Annotations

Cited by 11 publications

References 22 publications

Building Footprint Extraction in Dense Area from LiDAR Data using Mask R-CNN

Building Footprint Extraction in Dense Area from LiDAR Data using Mask R-CNN

A Review and Meta-analysis of Semantic Segmentation Models in Land Use/ Land Cover Mapping

Self-Filtered Learning for Semantic Segmentation of Buildings in Remote Sensing Imagery With Noisy Labels

Contact Info

Product

Resources

About