A new weakly supervised approach for ALS point cloud semantic segmentation

Although novel point cloud semantic segmentation schemes that continuously surpass state-of-the-art results exist, the success of learning an effective model typically relies on the availability of abundant labeled data. However, data annotation is a time-consumng and labor-intensive task, particula...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:ISPRS journal of photogrammetry and remote sensing Ročník 188; s. 237 - 254
Hlavní autoři: Wang, Puzuo, Yao, Wei
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 01.06.2022
Témata:
Consistency constraint
data collection
entropy
Entropy regularization
photogrammetry
Point cloud semantic segmentation
probability
Pseudo-label
Weakly supervised learning
Point cloud semantic segmentation
Consistency constraint
Weakly supervised learning
Pseudo-label
Entropy regularization
ISSN:0924-2716, 1872-8235
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Although novel point cloud semantic segmentation schemes that continuously surpass state-of-the-art results exist, the success of learning an effective model typically relies on the availability of abundant labeled data. However, data annotation is a time-consumng and labor-intensive task, particularly for large-scale airborne laser scanning (ALS) point clouds involving multiple classes in urban areas. Therefore, simultaneously obtaining promising results while significantly reducing labeling is crucial. In this study, we propose a deep-learning-based weakly supervised framework for the semantic segmentation of ALS point clouds. This is to exploit implicit information from unlabeled data subject to incomplete and sparse labels. Entropy regularization is introduced to penalize class overlap in the predictive probability. Additionally, a consistency constraint is designed to improve the robustness of the predictions by minimizing the difference between the current and ensemble predictions. Finally, we propose an online soft pseudo-labeling strategy to create additional supervisory sources in an efficient and nonparametric manner. Extensive experimental analysis using three benchmark datasets demonstrates that our proposed method significantly boosts the classification performance without compromising the computational efficiency, considering the sparse point annotations. It outperforms the current weakly supervised methods and achieves a result comparable to that of full supervision competitors. Considering the ISPRS Vaihingen 3D data, using only 1‰ labels, our method achieved an overall accuracy of 83.0% and an average F1 score of 70.0%. These increased by 6.9% and 12.8%, respectively, compared to the model trained only using sparse label information.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0924-2716
1872-8235
DOI:10.1016/j.isprsjprs.2022.04.016