ORIGINAL RESEARCH article
Front. Earth Sci.
Sec. Geohazards and Georisks
Volume 13 - 2025 | doi: 10.3389/feart.2025.1680510
This article is part of the Research TopicNatural Disaster Prediction Based on Experimental and Numerical MethodsView all 28 articles
Efficient vegetation filtering method using K-means clustering algorithm: case study of high steep rock slope
Provisionally accepted
- 1Shaoxing University, Shaoxing, China
- 2China University of Geosciences, Wuhan, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Rocky high steep slopes are prone to slope instability, resulting in serious disasters like landslides and avalanches. In practical engineering, the rock structural surface controls the slide stability. Accurate identification of the basic information of the rock structural surface is an important task for the effective management of high steep slopes. However, natural vegetation on the rock surface degrades its identification accuracy. In this paper, an innovative semi-automatic method with high precision is developed for filtering vegetation point clouds. The spatial location information and RGB color data are first subjected to dimensionality reduction using Principal Component Analysis (PCA) and the Red-Green Difference Index (RGDI). This step enables effective plane fitting for slope surfaces, excluding those resembling vertical plumb lines, and demonstrates strong applicability to steep slopes based on the PCA algorithm. Then, the K-means clustering algorithm is used for data segmentation, so that the point cloud data with similar location and color information form a cluster. Using a grid-based hierarchical subdivision method for selecting seed points can significantly improve computational efficiency. Finally, the clusters are progressively filtered by morphological features to remove redundant noise points. A preservation guarantee mechanism is established to prevent excessive filtering. The effectiveness of the filtering results is evaluated using class error Ie, IIe and total error Ae. Comparing with other algorithms, it is found that the class error Ie is 7.79%, the class error IIe is 4.34%, and the total result is limited to 6.53%. The proposed algorithm simplifies the structure and markedly enhances vegetation filtering accuracy, offering robust and practical support for structural surface identification on high-steep rock slopes.
Keywords: K-Means clustering, Three-dimensional point clouds, Vegetation removal, Filtering algorithm, Principal Component Analysis
Received: 06 Aug 2025; Accepted: 15 Sep 2025.
Copyright: © 2025 Guan, Yuan and Zheng. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Hongchao Zheng, zhenghongchao@cug.edu.cn
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.