# Individual Tree Segmentation from LiDAR Point Clouds for Urban Forest Inventory

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Study Area and Data

**Figure 2.**Study site shown in a color composite from the collected hyperspectral data. Measured tree locations are shown in green dots.

^{XT}GPS device in August 2010.

^{2}was obtained due to an intentional 80% overlap between flights. The raw LiDAR point cloud data were processed by TRSI using Microstation Terrasolid and TRSI proprietary software. Final products were provided in the commonly used LAS LiDAR format with a projection of NAD_1983_UTM_Zone_14N. Additionally, fine spatial resolution optical hyperspectral images were simultaneously acquired with the LiDAR data using an AISA Dual hyperspectral sensor. Hyperspectral images with 492 spectral bands were acquired at a spatial resolution of 1.6 m. Radiometric calibration and geometric correction were performed by the vendor. The original collected images were transformed from raw digital numbers to radiance by applying a pixel by pixel correction using the gain and offset. The offset was extracted from data stored in the “dark” files, which are sensor readings with a closed shutter in order to estimate sensor noise. The gain was extracted from the calibration files supplied by the sensor manufacturer. During the transformation from raw to radiance values, further corrections for dropped lines and spectral shift were also applied. A transformation from radiance to surface reflectance of the entire image dataset was then carried out using ATCOR, a Modtran-based code specifically developed to perform atmospheric correction. For the geometric correction, the vendor used the LiDAR positional information to compute the real world coordinates of each pixel in the dataset, followed by re-sampling of the hyperspectral images. The preprocessed images were then mosaicked and clipped for the study area. A false color composite from the hyperspectral data for the study site is shown in Figure 2.

## 3. Methodology

#### 3.1. A Framework to Combine LiDAR and Hyperspectral Data for Urban Forest Inventory

**Figure 3.**The designed framework to combine LiDAR and hyperspectral data for urban forest inventory at the individual tree level.

#### 3.2. LiDAR Data Filtering and Canopy Points Separation

#### 3.3. Treetop Detection

#### 3.4. Individual Tree Extraction

- (1)
- Randomly select one of the detected treetops p
_{t}(x_{t}, y_{t}, z_{t}) resulting from the treetop detection algorithm. - (2)
- Calculate the mean elevation of all points (x
_{i}, y_{i}, z_{i}) falling within a horizontal circle centered at the selected treetop (x_{t}, y_{t}, z_{t}):$$\overline{{Z}_{c}}=\frac{{\Sigma}_{i=1}^{N}{Z}_{i}}{N},if\sqrt{{\left({x}_{i}-{x}_{t}\right)}^{2}+{\left({y}_{i}-{y}_{t}\right)}^{2}}r$$_{i}, y_{i}, z_{i}) is the 3-D coordinate of each point, (x_{t}, y_{t}, z_{t}) is the 3-D coordinate of the circle center, and N is the total number of points within the circle. The initial radius (r) of the circle is the smallest crown size observed in the ground inventory. - (3)
- Compare the mean elevation ($\overline{{Z}_{c}}$) with the elevation of the selected treetop (Z
_{t}). If $\overline{{Z}_{c}}>{Z}_{t}$, reduce r to a smaller value (r-∆r) and repeat Steps 2 and 3 until $\overline{{Z}_{c}}<{Z}_{t}$, then export the radius of the circle as the estimated crown radius. Otherwise, for ($\overline{{Z}_{c}}<{Z}_{t}$) increase the radius of the circle by $\u2206r$ and get an expanded circle. Then calculate the mean elevation of the points falling within a horizontal donut ($\overline{{Z}_{d1}}$) defined by the previous circle and the expanded circle:$$\overline{{Z}_{d1}}=\frac{{\Sigma}_{i=1}^{N}{Z}_{i}}{N},if\sqrt{{\left({x}_{i}-{x}_{t}\right)}^{2}+{\left({y}_{i}-{y}_{t}\right)}^{2}}rand\sqrt{{\left({x}_{i}-{x}_{t}\right)}^{2}+{\left({y}_{i}-{y}_{t}\right)}^{2}}r+\u2206rand\overline{{Z}_{c}}{Z}_{t}$$ - (4)
- Compare $\overline{{Z}_{d1}}$ with $\overline{{Z}_{c}}$. If $\overline{{Z}_{d1}}>\overline{{Z}_{c}}$ or $\overline{{Z}_{d1}}$ = 0, indicating the donut expands to the boundary of trees, then export the radius of the circle as the estimated crown radius. Otherwise, for $\overline{{Z}_{d1}}<\overline{{Z}_{c}}$ increase the radius again by $\u2206r$ and get a newly expanded donut, which means the donut slides downward compared with its proceeding one. Then calculate the mean elevation of the points falling into the expanded new donut ($\overline{{Z}_{d2}}$) defined by the increased circle and its preceding circle:$$\overline{{Z}_{d2}}=\frac{{\Sigma}_{i=1}^{N}{Z}_{i}}{N},if\sqrt{{\left({x}_{i}-{x}_{t}\right)}^{2}+{\left({y}_{i}-{y}_{t}\right)}^{2}}r+\u2206rand\sqrt{{\left({x}_{i}-{x}_{t}\right)}^{2}+{\left({y}_{i}-{y}_{t}\right)}^{2}}r+2\u2206rand\overline{{Z}_{d1}}\overline{{Z}_{c}}$$
- (5)
- Compare $\overline{{Z}_{d2}}$ with $\overline{{Z}_{d1}}$. If $\overline{{Z}_{d2}}>\overline{{Z}_{d1}}$ or $\overline{{Z}_{d2}}$ = 0, export the radius of the proceeding donut as the estimated crown radius. Otherwise for $\overline{{Z}_{d2}}<\overline{{Z}_{d1}}$, meaning the donut slides again, repeat the process in Step 4 by increasing the radius r and getting an expended new donut and again calculating the mean elevation of the points falling into the expanded new donut.

## 4. Results and Discussion

#### 4.1. Filtering Results

**Figure 4.**(

**a**) Separated ground points in blue dots. (

**b**) Separated non-ground points in green dots for the study area.

#### 4.2. Detected Tree Numbers

**Table 1.**The number of LiDAR-detected, field-surveyed trees, and agreement of LiDAR detection for 17 selected testing subareas.

Subarea | No. of LiDAR-Detected Trees | No. of Field-Surveyed Trees | Agreement |
---|---|---|---|

1 | 69 | 68 | 98.5% |

2 | 196 | 186 | 94.6% |

3 | 117 | 116 | 99.1% |

4 | 45 | 42 | 92.9% |

5 | 235 | 242 | 97.1% |

6 | 502 | 488 | 97.1% |

7 | 97 | 93 | 95.7% |

8 | 215 | 227 | 94.8% |

9 | 131 | 128 | 97.7% |

10 | 41 | 32 | 71.9% |

11 | 191 | 163 | 82.8% |

12 | 161 | 170 | 94.8% |

13 | 68 | 72 | 94.4% |

14 | 52 | 48 | 91.7% |

15 | 41 | 46 | 90.1% |

16 | 263 | 257 | 97.7% |

17 | 228 | 226 | 99.1% |

Total | 2652 | 2604 | 93.5% |

#### 4.3. Tree Height Estimation

**Table 2.**Summary statistics of field-measured vs. LiDAR-estimated tree height, crown diameter, base height, and crown depth.

Field (meter) | LiDAR (meter) | |||||||
---|---|---|---|---|---|---|---|---|

Max. | Min. | Mean | Std. | Max. | Min. | Mean | Std. | |

Tree Height | 26.49 | 5.18 | 14.07 | 3.95 | 23.05 | 5.43 | 13.90 | 3.85 |

Crown Diameter | 28.02 | 4.12 | 13.37 | 5.84 | 25.96 | 3.86 | 11.70 | 4.75 |

Base Height | 8.32 | 1.49 | 4.13 | 1.20 | 8.50 | 1.12 | 4.54 | 1.53 |

Crown Depth | 21.61 | 2.38 | 9.86 | 3.51 | 21.36 | 2.25 | 8.93 | 3.25 |

**Table 3.**Statistics for regressions of field-measured and LiDAR-estimated tree height, crown diameter, base height, and crown depth with and without outliers.

With Outliers | Without Outliers | |||
---|---|---|---|---|

R^{2} | RMSE (meter) | R^{2} | RMSE (meter) | |

Tree Height | 0.93 | 1.11 | 0.98 | 0.57 |

Crown Diameter | 0.70 | 2.58 | 0.84 | 1.90 |

Base Height | 0.63 | 1.79 | 0.84 | 0.94 |

Crown Depth | 0.71 | 2.39 | 0.84 | 1.61 |

#### 4.4. Crown Size Estimation and Individual Tree Extraction

**Figure 6.**Detected trees shown in green dots for the study area (

**left**) and delineated crowns for a zoomed region (

**right**).

**Figure 7.**Defined crown boundary and extracted individual trees for a portion of the study area from (

**a**) the isotropic and (

**b**) anisotropic donut expanding and sliding method.

#### 4.5. Crown Base Height and Depth Estimation

## 5. Conclusions

## Acknowledgments

## Author Contributions

## Conflicts of Interest

## References

- Lim, K.; Treitz, P.; Wulder, M.; St-Onge, B.; Flood, M. LiDAR remote sensing of forest structure. Prog. Phys. Geogr.
**2003**, 27, 88–106. [Google Scholar] [CrossRef] - Hyyppä, J.; Hyyppä, H.; Leckie, D.; Gougeon, F.; Yu, X.; Maltamo, M. Review of methods of small-footprint airborne laser scanning for extracting forest inventory data in boreal forests. Int. J. Remote Sens.
**2008**, 29, 1339–1366. [Google Scholar] [CrossRef] - Hyyppä, J.; Inkinen, M. Detecting and estimating attributes for single trees using laser scanner. Photogramm. J. Finl.
**1999**, 16, 27–42. [Google Scholar] - Huang, Y.; Yu, B.; Zhou, J.; Hu, C.; Tan, W.; Hu, Z.; Wu, J. Toward automatic estimation of urban green volume using airborne LiDAR data and high resolution remote sensing images. Front. Earth Sci.
**2013**, 7, 43–54. [Google Scholar] [CrossRef] - Persson, Å.; Holmgren, J.; Söderman, U. Detecting and measuring individual trees using an airborne laser scanner. Photogramm. Eng. Remote Sens.
**2002**, 68, 925–932. [Google Scholar] - Brandtberg, T.; Warner, T.; Landenberger, R.; Mcgraw, J. Detection and analysis of individual leaf-off tree crowns in small footprint, high sampling density LiDAR data from the eastern deciduous forest in North America. Remote Sens. Environ.
**2003**, 85, 290–303. [Google Scholar] [CrossRef] - Leckie, D.; Gougeon, F.; Hill, D.; Quinn, R.; Armstrong, L.; Shreenan, R. Combined high-density lidar and multispectral imagery for individual tree crown analysis. Can. J. Remote Sens.
**2003**, 29, 1–17. [Google Scholar] [CrossRef] - Popescu, S.C.; Wynne, R.H. Seeing the trees in the forest: Using lidar and multispectral data fusion with local filtering and variable window size for estimating tree height. Photogramm. Eng. Remote Sens.
**2004**, 70, 589–604. [Google Scholar] [CrossRef] - Falkowski, M.J.; Smith, A.M.S.; Hudak, A.T.; Gessler, P.E.; Vierling, L.A.; Crookston, N.L. Automated estimation of individual conifer tree height and crown diameter via two-dimensional spatial wavelet analysis of LiDAR data. Can. J. Remote Sens.
**2006**, 32, 153–161. [Google Scholar] [CrossRef] - Solberg, S.; Naesset, E.; Bollandsas, O.M. Single tree segmentation using airborne laser scanner data in a structurally heterogeneous spruce forest. Photogramm. Eng. Remote Sens.
**2006**, 72, 1369–1378. [Google Scholar] [CrossRef] - Koch, B.; Heyder, U.; Weinacher, H. Detection of individual tree crowns in airborne Lidar data. Photogramm. Eng. Remote Sens.
**2006**, 72, 357–363. [Google Scholar] [CrossRef] - Chen, Q.; Baldocchi, D.; Gong, P.; Kelly, M. Isolating individual trees in a Savanna woodland using small footprint LiDAR data. Photogramm. Eng. Remote Sens.
**2006**, 72, 923–932. [Google Scholar] [CrossRef] - Kwak, D.; Lee, W.; Lee, J.; Biging, G.S.; Gong, P. Detection of individual trees and estimation of tree height using LiDAR data. J. For. Res.
**2007**, 12, 425–434. [Google Scholar] [CrossRef] - Hyyppä, J.; Yu, X.; Hyyppä, H.; Vastaranta, M.; Holopainen, M.; Kukko, A.; Kaartinen, H.; Jaakkola, A.; Vaaja, M.; Koskinen, J.; Alho, P. Advances in forest inventory using airborne laser scanning. Remote Sens.
**2012**, 4, 1190–1207. [Google Scholar] [CrossRef] - Kaartinen, H.; Hyyppä, J.; Yu, X.; Vastaranta, M.; Hyyppä, H.; Kukko, A.; Markus, H.; Heipke, C.; Hirschmugl, M.; Morsdorf, F.; et al. An international comparison of individual tree detection and extraction using airborne laser scanning. Remote Sens.
**2012**, 4, 950–974. [Google Scholar] [CrossRef] [Green Version] - Vauhkonen, J.; Ene, L.; Gupta, S.; Heinzel, J.; Holmgren, J.; Pitkänen, J.; Solberg, S.; Wang, Y.; Weinacker, H.; Hauglin, K.M.; et al. Comparative testing of single-tree detection algorithms under different types of forest. Forestry
**2012**, 85, 27–40. [Google Scholar] [CrossRef] - Jakubowski, M.K.; Li, W.; Guo, Q.; Kelly, M. Delineating individual trees from LiDAR data: A comparison of vector- and raster-based segmentation approaches. Remote Sens.
**2013**, 5, 4163–4186. [Google Scholar] [CrossRef] - Smith, S.L.; Holland, D.A.; Longley, P.A. The importance of understanding error in LiDAR digital elevation models. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci.
**2004**, 35, 996–1001. [Google Scholar] - Suárez, J.C.; Ontiveros, C.; Smith, S.; Snape, S. Use of airborne LiDAR and aerial photography in the estimation of individual tree heights in forestry. Comput. Geosci.
**2005**, 31, 253–262. [Google Scholar] [CrossRef] - Tiede, D.; Hochleitner, G.; Blaschke, T. A full GIS-based workflow for tree identification and tree crown delineation using laser scanning. In Proceedings of the ISPRS Workshop CMRT 2005: Object Extraction for 3D City Models, Road Databases and Traffic Monitoring—Concepts, Algorithms, and Evaluation, Vienna, Austria, 29–30 August 2005.
- Morsdorf, F.; Meier, E.; Kötz, B.; Itten, K.I.; Dobbertin, M.; Allgöwer, B. LiDAR-based geometric reconstruction of boreal type forest stands at single tree level for forest and wildland fire management. Remote Sens. Environ.
**2004**, 92, 353–362. [Google Scholar] [CrossRef] - Reitberger, J.; Schnörr, Cl.; Krzystek, P.; Stilla, U. 3D segmentation of single trees exploiting full waveform LiDAR data. ISPRS J. Photogramm. Remote Sens.
**2009**, 64, 561–574. [Google Scholar] [CrossRef] - Lee, H.; Slatton, K.C.; Roth, B.E.; Cropper, J.R.W.P. Adaptive clustering of airborne LiDAR data to segment individual tree crowns in managed pine forests. Int. J. Remote Sens.
**2010**, 31, 117–139. [Google Scholar] [CrossRef] - Tittmann, P.; Shafii, S.; Hartsough, B.; Hamann, B. Tree detection and delineation from LiDAR point clouds using RANSAC. In Proceedings of SilviLaser 2011, Hobart, TAS, Australia, 16–19 October 2011.
- Li, W.; Guo, Q.; Jakubowski, M.K.; Kelly, M. A new method for segmenting individual trees from the LiDAR point cloud. Photogramm. Eng. Remote Sens.
**2012**, 78, 75–84. [Google Scholar] [CrossRef] - Sima, M.C.; Nüchter, A. An extension of the Felzenszwalb-Huttenlocher segmentation to 3D point clouds. In Proceedings of 2012 5th International Conference on Machine Vision (ICMV 12), Wuhan, China, 20–21 October 2012.
- Haala, N.; Brenner, C. Extraction of buildings and trees in urban environments. ISPRS J. Photogramm. Remote Sens.
**1999**, 54, 130–137. [Google Scholar] [CrossRef] - Wu, B.; Yu, B.; Yue, W.; Shu, S.; Tan, W.; Hu, C.; Huang, Y.; Wu, J.; Liu, H. A voxel-based method for automated identification and morphological parameters estimation of individual street trees from mobile laser scanning data. Remote Sens.
**2013**, 5, 584–611. [Google Scholar] [CrossRef] - Liu, J.; Shen, J.; Zhao, R.; Xu, S. Extraction of individual tree crowns from airborne LiDAR data in human settlements. Math. Comput. Model.
**2013**, 58, 524–535. [Google Scholar] [CrossRef] - Holopainena, M.; Kankarea, V.; Vastarantaa, M.; Liang, X.; Lin, Y.; Vaajac, M.; Yu, X.; Hyyppä, J.; Hyyppä, H.; Kaartinen, H.; et al. Tree mapping using airborne, terrestrial and mobile laser scanning—A case study in heterogeneous urban forest. Urban For. Urban Green.
**2013**, 12, 546–553. [Google Scholar] [CrossRef] - Saarinen, N.; Vastaranta, M.; Kankare, V.; Tanhuanpää, T.; Holopainen, M.; Hyyppä, J.; Hyyppä, H. Urban-tree-attribute update using multisource single-tree inventory. Forests
**2014**, 5, 1032–1052. [Google Scholar] [CrossRef] - Secord, J.; Zakhor, A. Tree detection in urban regions using aerial LiDAR and image data. IEEE Geosci. Remote Sens. Lett.
**2007**, 4, 196–200. [Google Scholar] [CrossRef] - Kim, J.; Muller, J.-P. Tree and building detection in dense urban environments using automated processing of IKONOS image and LiDAR data. Int. J. Remote Sens.
**2011**, 32, 2245–2273. [Google Scholar] [CrossRef] - Lafarge, E.; Mallet, C. Creating large-scale city models from 3D-point clouds: A robust approach with hybrid representation. Int. J. Comput. Vis.
**2012**, 99, 69–85. [Google Scholar] [CrossRef] - Zhang, C.; Qiu, F. Mapping individual tree species for an urban forest using airborne LiDAR and hyperspectral imagery. Photogramm. Eng. Remote Sens.
**2012**, 78, 1079–1087. [Google Scholar] [CrossRef] - Sithole, G.; Vosselman, G. Experimental comparison of filter algorithms for bare-earth extraction from airborne laser scanning point clouds. ISPRS J. Photogramm. Remote Sens.
**2004**, 59, 85–101. [Google Scholar] [CrossRef] - Chang, J. Segmentation-Based Filtering and Object-Based Feature Extraction from Airborne LiDAR data. Ph.D. Thesis, University of Texas at Dallas, Richardson, TX, USA, 2011. [Google Scholar]
- Gill, S.J.; Biging, G.S.; Murphy, E.C. Modeling conifer tree crown radius and estimating canopy cover. For. Ecol. Manag.
**2000**, 126, 405–416. [Google Scholar] [CrossRef] - Popescu, S.C.; Wynne, R.H.; Nelson, R.F. Measuring individual tree crown diameter with LiDAR and assessing its influence on estimating forest volume and biomass. Can. J. Remote Sens.
**2003**, 29, 564–577. [Google Scholar] [CrossRef] - Heurich, M.; Persson, Å.; Holmgren, J.; Kennel, E. Detecting and measuring individual trees with laser scanning in mixed mountain forest of central Europe using an algorithm developed for Swedish boreal forest conditions. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci.
**2004**, XXXVI-8/W2, 307–312. [Google Scholar]

© 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Zhang, C.; Zhou, Y.; Qiu, F.
Individual Tree Segmentation from LiDAR Point Clouds for Urban Forest Inventory. *Remote Sens.* **2015**, *7*, 7892-7913.
https://doi.org/10.3390/rs70607892

**AMA Style**

Zhang C, Zhou Y, Qiu F.
Individual Tree Segmentation from LiDAR Point Clouds for Urban Forest Inventory. *Remote Sensing*. 2015; 7(6):7892-7913.
https://doi.org/10.3390/rs70607892

**Chicago/Turabian Style**

Zhang, Caiyun, Yuhong Zhou, and Fang Qiu.
2015. "Individual Tree Segmentation from LiDAR Point Clouds for Urban Forest Inventory" *Remote Sensing* 7, no. 6: 7892-7913.
https://doi.org/10.3390/rs70607892