4.7 Article

Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds

Journal

REMOTE SENSING
Volume 11, Issue 17, Pages -

Publisher

MDPI
DOI: 10.3390/rs11171961

Keywords

airborne laser scanning; power corridor classification; class distribution; feature selection; random forest

Funding

  1. National Natural Science Foundation of China [41871264]
  2. Youth Innovation Promotion Association CAS
  3. Major Projects of High-Resolution Earth Observation (Civil Part) [30-Y20A15-9003-17/18]

Ask authors/readers for more resources

Power corridor classification using LiDAR (light detection and ranging) point clouds is an important means for power line inspection. Many supervised classification methods have been used for classifying power corridor scenes, such as using random forest (RF) and JointBoost. However, these studies did not systematically analyze all the relevant factors that affect the classification, including the class distribution, feature selection, classifier type and neighborhood radius for classification feature extraction. In this study, we examine these factors using point clouds collected by an airborne laser scanning system (ALS). Random forest shows strong robustness to various pylon types. When classifying complex scenes, the gradient boosting decision tree (GBDT) shows good generalization. Synthetically, considering performance and efficiency, RF is very suitable for power corridor classification. This study shows that balanced learning leads to poor classification performance in the current scene. Data resampling for the original unbalanced dataset may not be necessary. The sensitivity analysis shows that the optimal neighborhood radius for feature extraction of different objects may be different. Scale invariance and automatic scale selection methods should be further studied. Finally, it is suggested that RF, original unbalanced class distribution, and complete feature set should be considered for power corridor classification in most cases.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available