Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds
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 d...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2019-08-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/11/17/1961 |
id |
doaj-96f4b9843c1c42cba1fd43d1ae06ce2e |
---|---|
record_format |
Article |
spelling |
doaj-96f4b9843c1c42cba1fd43d1ae06ce2e2020-11-25T01:12:22ZengMDPI AGRemote Sensing2072-42922019-08-011117196110.3390/rs11171961rs11171961Systematic Comparison of Power Corridor Classification Methods from ALS Point CloudsShuwen Peng0Xiaohuan Xi1Cheng Wang2Pinliang Dong3Pu Wang4Sheng Nie5Key Lab of Digital Earth, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, ChinaKey Lab of Digital Earth, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, ChinaKey Lab of Digital Earth, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, ChinaDepartment of Geography and the Environment, University of North Texas, Denton, TX 76203, USAKey Lab of Digital Earth, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, ChinaKey Lab of Digital Earth, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, ChinaPower 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.https://www.mdpi.com/2072-4292/11/17/1961airborne laser scanningpower corridor classificationclass distributionfeature selectionrandom forest |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Shuwen Peng Xiaohuan Xi Cheng Wang Pinliang Dong Pu Wang Sheng Nie |
spellingShingle |
Shuwen Peng Xiaohuan Xi Cheng Wang Pinliang Dong Pu Wang Sheng Nie Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds Remote Sensing airborne laser scanning power corridor classification class distribution feature selection random forest |
author_facet |
Shuwen Peng Xiaohuan Xi Cheng Wang Pinliang Dong Pu Wang Sheng Nie |
author_sort |
Shuwen Peng |
title |
Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds |
title_short |
Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds |
title_full |
Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds |
title_fullStr |
Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds |
title_full_unstemmed |
Systematic Comparison of Power Corridor Classification Methods from ALS Point Clouds |
title_sort |
systematic comparison of power corridor classification methods from als point clouds |
publisher |
MDPI AG |
series |
Remote Sensing |
issn |
2072-4292 |
publishDate |
2019-08-01 |
description |
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. |
topic |
airborne laser scanning power corridor classification class distribution feature selection random forest |
url |
https://www.mdpi.com/2072-4292/11/17/1961 |
work_keys_str_mv |
AT shuwenpeng systematiccomparisonofpowercorridorclassificationmethodsfromalspointclouds AT xiaohuanxi systematiccomparisonofpowercorridorclassificationmethodsfromalspointclouds AT chengwang systematiccomparisonofpowercorridorclassificationmethodsfromalspointclouds AT pinliangdong systematiccomparisonofpowercorridorclassificationmethodsfromalspointclouds AT puwang systematiccomparisonofpowercorridorclassificationmethodsfromalspointclouds AT shengnie systematiccomparisonofpowercorridorclassificationmethodsfromalspointclouds |
_version_ |
1725166813401055232 |