Comparison of UAV-based LiDAR and Digital Aerial Photogrammetry for Measuring Crown-Level Canopy Height in the Urban Environment
Longfei Zhou,Ran Meng,Yiyang Tan,Zhengang Lv,Yutao Zhao,Binyuan Xu,Feng Zhao
DOI: https://doi.org/10.1016/j.ufug.2022.127489
IF: 5.766
2022-01-01
Urban Forestry & Urban Greening
Abstract:Spatial information on urban forest canopy height (FCH) is fundamental for urban forest monitoring and assisting urban planning and management. Traditionally, ground-based canopy height measurements are time-consuming and laborious, making it challenging for periodic inventory of urban FCH at crown level. Airborne-light detection and ranging (LiDAR) sensor can efficiently measure crown-level FCH; however, the high cost of airborne-LiDAR data collection over large scales hinders its wide applications at a high temporal resolution. Previous studies have shown that in some cases, the Unmanned Aerial Vehicle (UAV)-digital aerial photogrammetry (DAP) approach (i.e., UAV-based structure from motion algorithm) is equivalent to or even outperform airborne-LiDAR in measuring forest structure, but few studies have evaluated their performances in measuring FCH in more complex urban environment, across non-ground coverage (including both canopy and building coverage) and topographical slope gradients. Also, the contribution of multi-angle measurement technique from UAV-DAP to FCH estimation accuracy has rarely been explored in the urban environment. Here, we compared the performances of UAV-LiDAR and UAV-DAP approaches on measuring thousands of crown-level FCH at different non-ground coverage and topographical slope areas in an urban environment. Specifically, UAV-LiDAR-based spatial measurements of crown-level FCH were used as the reference after ground-based validation (R2 = 0.88, RMSE = 2.36 m). The accuracy of UAV-DAP approach with/without multi-angle measurement in different non-ground coverage and topographical slope areas were then analyzed. The results showed that although the DAP multi-angle-based approach can improve the accuracy of spatial measurement for crown-level FCH in some cases, non-ground coverage (including both canopy and building coverage) was still the main factor affecting the broad applications of DAP approach in measuring urban FCH: at areas where non-ground coverage < 0.95, no matter how topographical slope varied, the accuracy of DAP approach was high (R2 = 0.86∼0.94, RMSE = 1.56∼2.93 m); at areas where non-ground coverage > 0.95, except for the case of flat areas (i.e., topographical slope <= 2°), the accuracy of DAP was poor (R2 = 0.20, RMSE = 12.34 m). However, using LiDAR-digital terrain model (DTM) instead of DAP-DTM, at areas where non-ground coverage > 0.95, can significantly improve the accuracy of UAV-DAP approach in measuring crown-level FCH (R2 = 0.91, RMSE =1.61 m). Our study thus provides a complete guidance on the usage of cost-effective UAV-DAP approach for measuring crown-level FCH in the urban environment, which will be helpful for precise urban forest management and improving the efficiency of urban environmental planning.