# Estimating Individual Tree Above-Ground Biomass of Chinese Fir Plantation: Exploring the Combination of Multi-Dimensional Features from UAV Oblique Photos

^{1}

^{2}

^{*}

^{†}

## Abstract

**:**

^{2}= 0.79; RMSECV = 44.77 kg; and rRMSECV = 0.25). Comparing the performance of estimating IT-AGB models with different spatial resolution images (0.05, 0.1, 0.2, 0.5 and 1 m), the model was the best at the spatial resolution of 0.2 m, which was significantly better than that of the other four. Moreover, we also divided the individual tree canopy into four directions (East, West, South and North) to develop estimation models respectively. The result showed that the IT-AGB estimation capacity varied significantly in different directions, and the West-model had better performance, with the estimation accuracy of 67%. This study indicates the potential of using oblique photogrammetry technology to estimate AGB at an individual tree scale, which can support carbon stock estimation as well as precision forestry application.

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Study Area

^{2}. The forest farm belongs to subtropical monsoon climate, with sufficient sunshine and rainfall. The annual average temperature is 21.7 °C and the annual average precipitation is 1300 mm. The elevation varies from 125 m to 300 m. The forest type is mainly artificial forest with tall and dense trees, high canopy density and rich tree species. The dominant tree species include Chinese fir (Cunninghamia lanceolata (Lamb.) Hook.), eucalyptus (Eucalyptus robusta Smith), Castanopsis hystrix Miq. (Castanopsis hystrix J. D. Hooker et Thomson ex A. De Candolle). Among them, Chinese fir has ecological, economic and medicinal values. However, the complex terrain and high canopy density of the forest farm bring challenges to the estimation of IT-AGB.

#### 2.2. Data

#### 2.2.1. Field Data

#### 2.2.2. UAV Oblique Photography Data and Auxiliary Terrain Data

#### 2.3. Data Processing

#### 2.4. Methods

#### 2.4.1. Individual Tree Segmentation

#### 2.4.2. Feature Variables Extraction

#### 2.4.3. Construction of Empirical Model

#### 2.4.4. Accuracy Verification

^{2}), root mean square error of leave-one-out cross-validation (RMSECV) and relative RMSECV (rRMSECV) to verify the predictive ability of the models.

## 3. Results

#### 3.1. IT-AGB Distribution of Sample Plots

#### 3.2. Features Selection

_{5}, height variables H

_{max}and H

_{30th}, and intensity variables sum

_{sqrt_i}and I

_{SUM_SQRT}. GLA is the color vegetation index, which can be used to distinguish vegetation from non-vegetation. Point cloud height variables reflects forest vertical structure information. Point cloud density variables is highly sensitive to forest AGB, and reflects forest horizontal structure information. The intensity value calculated based on RGB and HSI color space has strong correlation with IT-AGB (Figure 6).

#### 3.3. Contribution of Different Intensity Information and Spatial Resolution on Model

^{2}and RMSECV were 0.34~0.79 and 44.77~82.24 kg respectively, and rRMSECV was 0.25~0.45. The estimation accuracy of the model was the best at a spatial resolution of 0.2 m, which was significantly improved compared with the spatial resolution of the other four. However, when the spatial resolution was increased from 1 m to 0.5 m, both the value of RMSECV or rRMSECV were similar or same, which meant that the model accuracy was not significantly improved.

^{2}(0.59, 0.72, 0.79, 0.74 and 0.58) and the lowest RMSECV (61.41, 50.10, 44.77, 49.45 and 66.74 kg) at corresponding resolutions (0.05, 0.1, 0.2, 0.5 and 1 m). Taking the spatial resolution of 0.2 m as an example, compared with the intensity variable of RGB space, the contribution of HSI space variable to the estimation model was slightly lower, with R

^{2}of 0.75 and RMSECV of 48.52 kg. The R

^{2}of the experiment A models was low, only 0.50. Overall, the IT-AGB estimation model of experiment C was more stable and robust.

#### 3.4. Comparisons of Models in Different Tree Canopy Directions

^{2}and rRMSECV of the estimation models in different directions ranged from 0.29~0.60 and 0.33~0.45 respectively. The results showed that the performance of West-model was greater than that of South-model, North-model and East-model, and the estimation accuracy was improved by 2%, 8% and 12% respectively. The West-model can best reflect the overall IT-AGB with R

^{2}reaching 0.60 and RMSECV of 60.17 kg. The second was the South-model with R

^{2}of 0.58 and rRMSECV of 0.35. The East-model was the worst with R

^{2}of only 0.29 and RMSECV of 80.79 kg.

## 4. Discussion

#### 4.1. Extraction of Feature Variables for Modeling

_{5}, H

_{max}, H

_{30th}, sum

_{sqrt_i}and I

_{SUM_SQRT}were selected as the final modeling variables by Pearson correlation analysis and stepwise regression. The Chinese fir has lanceolate or strip lanceolate leaves, conical crown and concentrated crown distribution. The point cloud density can digitize the above features and better highlight the forest horizontal structure crown information. The H

_{max}and H

_{30th}are point cloud height variables that reflect the information of height related to the forest AGB and the vertical structure of the forest crown [25,60,61]. The GLA can effectively distinguish vegetation from other features, and is a good correlation between vegetation index and AGB [38,41]. Furthermore, the sum

_{sqrt_i}(intensity variable of the RGB spatial) [42] and I

_{SUM_SQRT}(intensity variable of HSI spatial) can reflect image features and the spatial information contained in the image. The relatively compact canopy structure rarely allows ground information to pass through, so there are fewer mutations in intensity information. This makes the description of the forest canopy surface information more accurate, which can explain the usefulness of the intensity information to IT-AGB. In brief, the combination of horizontal and vertical structure information with RGB or HSI spatial intensity variables can reflect the IT-AGB information more comprehensively.

#### 4.2. Effects of Different Spatial Resolution on Models

#### 4.3. Effects of Intensity Information on Model

#### 4.4. Effects of Different Direction on Model

_{1}and T

_{2}(or T

_{3}) compete for conditions such as space and sunlight in the growth process. Among them, the competition between T

_{1}and T

_{3}was essentially a competition between the westward canopy of T

_{1}and the eastward canopy of T

_{3}, and the obvious advantage of T

_{1}was mainly manifested in the relatively high altitude. Similarly, the relationship between T

_{1}and T

_{2}was the competition between the northward crown of T

_{1}and the southward crown of T

_{2}, which was a competition of equal status. This can also explain that the West-model performs better than that of South-model in the canopy of T

_{1}. This could also explain the superior performance of the West-model in the T

_{1}canopy compared to the South-model.

## 5. Conclusions

## Author Contributions

## Funding

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## References

- Gómez, C.; White, J.C.; Wulder, M.A.; Alejandro, P. Historical forest biomass dynamics modelled with Landsat spectral trajectories. ISPRS J. Photogramm. Remote Sens.
**2014**, 93, 14–28. [Google Scholar] [CrossRef] [Green Version] - Zhu, X.; Liu, D. Improving forest aboveground biomass estimation using seasonal Landsat NDVI time-series. ISPRS J. Photogramm. Remote Sens.
**2015**, 102, 222–231. [Google Scholar] [CrossRef] - Li, L.; Guo, Q.; Tao, S.; Kelly, M.; Xu, G. Lidar with multi-temporal MODIS provide a means to upscale predictions of forest biomass. ISPRS J. Photogramm. Remote Sens.
**2015**, 102, 198–208. [Google Scholar] [CrossRef] - Luo, S.; Wang, C.; Xi, X.; Nie, S.; Fan, X.; Chen, H.; Ma, D.; Liu, J.; Zou, J.; Lin, Y.; et al. Estimating forest aboveground biomass using small-footprint full-waveform airborne LiDAR data. Int. J. Appl. Earth Obs. Geoinf.
**2019**, 83, 101922. [Google Scholar] [CrossRef] - Malhi, R.K.M.; Anand, A.; Srivastava, P.K.; Chaudhary, S.K.; Pandey, M.K.; Behera, M.D.; Kumar, A.; Singh, P.; Sandhya Kiran, G. Synergistic evaluation of Sentinel 1 and 2 for biomass estimation in a tropical forest of India. Adv. Space Res.
**2021**, in press. [Google Scholar] [CrossRef] - Campbell, M.J.; Dennison, P.E.; Kerr, K.L.; Brewer, S.C.; Anderegg, W.R.L. Scaled biomass estimation in woodland ecosystems: Testing the individual and combined capacities of satellite multispectral and lidar data. Remote Sens. Environ.
**2021**, 262, 112511. [Google Scholar] [CrossRef] - Wang, G.; Zhang, M.; Gertner, G.Z.; Oyana, T.; McRoberts, R.E.; Ge, H. Uncertainties of mapping aboveground forest carbon due to plot locations using national forest inventory plot and remotely sensed data. Scand. J. For. Res.
**2011**, 26, 360–373. [Google Scholar] [CrossRef] - Wang, Y.; Ni, W.; Sun, G.; Chi, H.; Zhang, Z.; Guo, Z. Slope-adaptive waveform metrics of large footprint lidar for estimation of forest aboveground biomass. Remote Sens. Environ.
**2019**, 224, 386–400. [Google Scholar] [CrossRef] - Wang, Y.; Zhang, X.; Guo, Z. Estimation of tree height and aboveground biomass of coniferous forests in North China using stereo ZY-3, multispectral Sentinel-2, and DEM data. Ecol. Indic.
**2021**, 126, 107645. [Google Scholar] [CrossRef] - Liao, Z.; He, B.; Quan, X.; van Dijk, A.I.J.M.; Qiu, S.; Yin, C. Biomass estimation in dense tropical forest using multiple information from single-baseline P-band PolInSAR data. Remote Sens. Environ.
**2019**, 221, 489–507. [Google Scholar] [CrossRef] - Ploton, P.; Barbier, N.; Couteron, P.; Antin, C.M.; Ayyappan, N.; Balachandran, N.; Barathan, N.; Bastin, J.F.; Chuyong, G.; Dauby, G.; et al. Toward a general tropical forest biomass prediction model from very high resolution optical satellite images. Remote Sens. Environ.
**2017**, 200, 140–153. [Google Scholar] [CrossRef] - Askne, J.I.H.; Soja, M.J.; Ulander, L.M.H. Biomass estimation in a boreal forest from TanDEM-X data, lidar DTM, and the interferometric water cloud model. Remote Sens. Environ.
**2017**, 196, 265–278. [Google Scholar] [CrossRef] - Yang, H.; Li, F.; Wang, W.; Yu, K. Estimating Above-Ground Biomass of Potato Using Random Forest and Optimized Hyperspectral Indices. Remote Sens.
**2021**, 13, 2339. [Google Scholar] [CrossRef] - Proisy, C.; Couteron, P.; Fromard, F.O. Predicting and mapping mangrove biomass from canopy grain analysis using Fourier-based textural ordination (FOTO) of IKONOS images. Remote Sens. Environ.
**2007**, 109, 379–392. [Google Scholar] [CrossRef] - Yadav, S.; Padalia, H.; Sinha, S.K.; Srinet, R.; Chauhan, P. Above-ground biomass estimation of Indian tropical forests using X band Pol-InSAR and Random Forest. Remote Sens. Appl. Soc. Environ.
**2021**, 21, 100462. [Google Scholar] [CrossRef] - Li, G.; Xie, Z.; Jiang, X.; Lu, D.; Chen, E. Integration of ZiYuan-3 Multispectral and Stereo Data for Modeling Aboveground Biomass of Larch Plantations in North China. Remote Sens.
**2019**, 11, 2328. [Google Scholar] [CrossRef] [Green Version] - Zhao, P.; Lu, D.; Wang, G.; Wu, C.; Huang, Y.; Yu, S. Examining Spectral Reflectance Saturation in Landsat Imagery and Corresponding Solutions to Improve Forest Aboveground Biomass Estimation. Remote Sens.
**2016**, 8, 469. [Google Scholar] [CrossRef] [Green Version] - Hu, T.; Zhang, Y.; Su, Y.; Zheng, Y.; Lin, G.; Guo, Q. Mapping the Global Mangrove Forest Aboveground Biomass Using Multisource Remote Sensing Data. Remote Sens.
**2020**, 12, 1690. [Google Scholar] [CrossRef] - Cao, L.; Coops, N.C.; Innes, J.L.; Sheppard, S.R.J.; Fu, L.; Ruan, H.; She, G. Estimation of forest biomass dynamics in subtropical forests using multi-temporal airborne LiDAR data. Remote Sens. Environ.
**2016**, 178, 158–171. [Google Scholar] [CrossRef] - Bazezew, M.N.; Hussin, Y.A.; Kloosterman, E.H. Integrating Airborne LiDAR and Terrestrial Laser Scanner forest parameters for accurate above-ground biomass/carbon estimation in Ayer Hitam tropical forest, Malaysia. Int. J. Appl. Earth Obs. Geoinf.
**2018**, 73, 638–652. [Google Scholar] [CrossRef] - Duncanson, L.I.; Dubayah, R.O.; Cook, B.D.; Rosette, J.; Parker, G. The importance of spatial detail: Assessing the utility of individual crown information and scaling approaches for lidar-based biomass density estimation. Remote Sens. Environ.
**2015**, 168, 102–112. [Google Scholar] [CrossRef] - Lu, J.; Wang, H.; Qin, S.; Cao, L.; Pu, R.; Li, G.; Sun, J. Estimation of aboveground biomass of Robinia pseudoacacia forest in the Yellow River Delta based on UAV and Backpack LiDAR point clouds. Int. J. Appl. Earth Obs. Geoinf.
**2020**, 86, 102014. [Google Scholar] [CrossRef] - Nord-Larsen, T.; Schumacher, J. Estimation of forest resources from a country wide laser scanning survey and national forest inventory data. Remote Sens. Environ.
**2012**, 119, 148–157. [Google Scholar] [CrossRef] - Eitel, J.U.H.; Höfle, B.; Vierling, L.A.; Abellán, A.; Asner, G.P.; Deems, J.S.; Glennie, C.L.; Joerg, P.C.; LeWinter, A.L.; Magney, T.S.; et al. Beyond 3-D: The new spectrum of lidar applications for earth and ecological sciences. Remote Sens. Environ.
**2016**, 186, 372–392. [Google Scholar] [CrossRef] [Green Version] - Su, Y.; Guo, Q.; Xue, B.; Hu, T.; Alvarez, O.; Tao, S.; Fang, J. Spatial distribution of forest aboveground biomass in China: Estimation through combination of spaceborne lidar, optical imagery, and forest inventory data. Remote Sens. Environ.
**2016**, 173, 187–199. [Google Scholar] [CrossRef] [Green Version] - Liu, K.; Shen, X.; Cao, L.; Wang, G.; Cao, F. Estimating forest structural attributes using UAV-LiDAR data in Ginkgo plantations. ISPRS J. Photogramm. Remote Sens.
**2018**, 146, 465–482. [Google Scholar] [CrossRef] - Jayathunga, S.; Owari, T.; Tsuyuki, S. The use of fixed–wing UAV photogrammetry with LiDAR DTM to estimate merchantable volume and carbon stock in living biomass over a mixed conifer–broadleaf forest. Int. J. Appl. Earth Obs. Geoinf.
**2018**, 73, 767–777. [Google Scholar] [CrossRef] - Li, B.; Xu, X.; Zhang, L.; Han, J.; Bian, C.; Li, G.; Liu, J.; Jin, L. Above-ground biomass estimation and yield prediction in potato by using UAV-based RGB and hyperspectral imaging. ISPRS J. Photogramm. Remote Sens.
**2020**, 162, 161–172. [Google Scholar] [CrossRef] - Fu, X.; Zhang, Z.; Cao, L.; Coops, N.C.; Goodbody, T.R.H.; Liu, H.; Shen, X.; Wu, X. Assessment of approaches for monitoring forest structure dynamics using bi-temporal digital aerial photogrammetry point clouds. Remote Sens. Environ.
**2021**, 255, 112300. [Google Scholar] [CrossRef] - Dandois, J.P.; Ellis, E.C. Remote Sensing of Vegetation Structure Using Computer Vision. Remote Sens.
**2010**, 2, 1157–1176. [Google Scholar] [CrossRef] [Green Version] - Goodbody, T.R.H.; Coops, N.C.; White, J.C. Digital Aerial Photogrammetry for Updating Area-Based Forest Inventories: A Review of Opportunities, Challenges, and Future Directions. Curr. For. Rep.
**2019**, 5, 55–75. [Google Scholar] [CrossRef] [Green Version] - Pulkkinen, M.; Ginzler, C.; Traub, B.; Lanz, A. Stereo-imagery-based post-stratification by regression-tree modelling in Swiss National Forest Inventory. Remote Sens. Environ.
**2018**, 213, 182–194. [Google Scholar] [CrossRef] - Watts, A.C.; Ambrosia, V.G.; Hinkley, E.A. Unmanned Aircraft Systems in Remote Sensing and Scientific Research: Classification and Considerations of Use. Remote Sens.
**2012**, 4, 1671–1692. [Google Scholar] [CrossRef] [Green Version] - Puliti, S.; Ene, L.T.; Gobakken, T.; Næsset, E. Use of partial-coverage UAV data in sampling for large scale forest inventories. Remote Sens. Environ.
**2017**, 194, 115–126. [Google Scholar] [CrossRef] - Dandois, J.P.; Ellis, E.C. High spatial resolution three-dimensional mapping of vegetation spectral dynamics using computer vision. Remote Sens. Environ.
**2013**, 136, 259–276. [Google Scholar] [CrossRef] [Green Version] - Cunliffe, A.M.; Brazier, R.E.; Anderson, K. Ultra-fine grain landscape-scale quantification of dryland vegetation structure with drone-acquired structure-from-motion photogrammetry. Remote Sens. Environ.
**2016**, 183, 129–143. [Google Scholar] [CrossRef] [Green Version] - Price, B.; Waser, L.T.; Wang, Z.; Marty, M.; Ginzler, C.; Zellweger, F. Predicting biomass dynamics at the national extent from digital aerial photogrammetry. Int. J. Appl. Earth Obs. Geoinf.
**2020**, 90, 102116. [Google Scholar] [CrossRef] - Alonzo, M.; Dial, R.J.; Schulz, B.K.; Andersen, H.-E.; Lewis-Clark, E.; Cook, B.D.; Morton, D.C. Mapping tall shrub biomass in Alaska at landscape scale using structure-from-motion photogrammetry and lidar. Remote Sens. Environ.
**2020**, 245, 111841. [Google Scholar] [CrossRef] - Luo, Y.; Lu, F. Comprehensive Database of Biomes Regressions for China’s Tree Species; China Forestry Publishing House: Beijing, China, 2015; p. 125. [Google Scholar]
- Michez, A.; Piegay, H.; Lisein, J.; Claessens, H.; Lejeune, P. Classification of riparian forest species and health condition using multi-temporal and hyperspatial imagery from unmanned aerial system. Environ. Monit. Assess
**2016**, 188, 146. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Jing, R.; Gong, Z.; Zhao, W.; Pu, R.; Deng, L. Above-bottom biomass retrieval of aquatic plants with regression models and SfM data acquired by a UAV platform—A case study in Wild Duck Lake Wetland, Beijing, China. ISPRS J. Photogramm. Remote Sens.
**2017**, 134, 122–134. [Google Scholar] [CrossRef] - Giannetti, F.; Chirici, G.; Gobakken, T.; Næsset, E.; Travaglini, D.; Puliti, S. A new approach with DTM-independent metrics for forest growing stock prediction using UAV photogrammetric data. Remote Sens. Environ.
**2018**, 213, 195–205. [Google Scholar] [CrossRef] - Navarro, A.; Young, M.; Allan, B.; Carnell, P.; Macreadie, P.; Ierodiaconou, D. The application of Unmanned Aerial Vehicles (UAVs) to estimate above-ground biomass of mangrove ecosystems. Remote Sens. Environ.
**2020**, 242, 111747. [Google Scholar] [CrossRef] - Puliti, S.; Ørka, H.; Gobakken, T.; Næsset, E. Inventory of Small Forest Areas Using an Unmanned Aerial System. Remote Sens.
**2015**, 7, 9632–9654. [Google Scholar] [CrossRef] [Green Version] - Dalponte, M.; Ørka, H.O.; Ene, L.T.; Gobakken, T.; Næsset, E. Tree crown delineation and tree species classification in boreal forests using hyperspectral and ALS data. Remote Sens. Environ.
**2014**, 140, 306–317. [Google Scholar] [CrossRef] - Hyyppa, J.; Kelle, M.; Lehikoinen, M.; Inkinen, M. A segmentation-based method to retrieve stem volume estimates from 3-D tree height models produced by laser scanners. IEEE Trans. Geosci. Remote Sens.
**2001**, 39, 969–975. [Google Scholar] [CrossRef] - Lee, H.; Slatton, K.C.; Roth, B.E.; Cropper, 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] - Vastaranta, M. Individual tree detection and area-based approach in retrieval of forest inventory characteristics from low-pulse airborne laser scanning data. Photogramm. J. Finl.
**2011**, 22, 1311–1317. [Google Scholar] - Hu, B.; Li, J.; Jing, L.; Judah, A. Improving the efficiency and accuracy of individual tree crown delineation from high-density LiDAR data. Int. J. Appl. Earth Obs. Geoinf.
**2014**, 26, 145–155. [Google Scholar] [CrossRef] - Mongus, D.; Žalik, B. An efficient approach to 3D single tree-crown delineation in LiDAR data. ISPRS J. Photogramm. Remote Sens.
**2015**, 108, 219–233. [Google Scholar] [CrossRef] - Hu, X.; Chen, W.; Xu, W. Adaptive Mean Shift-Based Identification of Individual Trees Using Airborne LiDAR Data. Remote Sens.
**2017**, 9, 148. [Google Scholar] [CrossRef] [Green Version] - Dai, W.; Yang, B.; Dong, Z.; Shaker, A. A new method for 3D individual tree extraction using multispectral airborne LiDAR point clouds. ISPRS J. Photogramm. Remote Sens.
**2018**, 144, 400–411. [Google Scholar] [CrossRef] - Wu, Y.; Zhao, L.; Jiang, H.; Guo, X.; Huang, F. Image segmentation method for green crops using improved mean shift. Trans. Chin. Soc. Agric. Eng.
**2014**, 30, 161–167. [Google Scholar] - Jing, R.; Deng, L.; Zhao, W.J.; Gong, Z.N. Object-oriented aquatic vegetation extracting approach based on visible vegetation indices. Chin. J. Appl. Ecol.
**2016**, 27, 1427–1436. [Google Scholar] - Sun, G.; Wang, X.; Yan, T.; Li, X.; Chen, J. Inversion method of flora growth parameters based on machine vision. Trans. Chin. Soc. Agric. Eng.
**2014**, 30, 187–195. [Google Scholar] - Gitelson, A.A.; Kaufman, Y.J.; Stark, R.; Rundquist, D. Novel algorithms for remote estimation of vegetation fraction. Remote Sens. Environ.
**2002**, 80, 76–87. [Google Scholar] [CrossRef] [Green Version] - Hague, T.; Tillett, N.D.; Wheeler, H. Automated Crop and Weed Monitoring in Widely Spaced Cereals. Precis. Agric.
**2006**, 7, 21–32. [Google Scholar] [CrossRef] - Montalvo, M.; Guerrero, J.M.; Romeo, J.; Emmi, L.; Guijarro, M.; Pajares, G. Automatic expert system for weeds/crops identification in images from maize fields. Expert Syst. Appl.
**2013**, 40, 75–82. [Google Scholar] [CrossRef] - Wu, Y.; Zhang, H.; Lu, T.; Xu, H.; Ou, G. Aboveground biomass estimation of coniferous forests using Landsat 8 OLI and the determination of saturation points in the alpine and sub-alpine area of Northwest Yunnan. J. Yunnan Univ. Nat. Sci. Ed.
**2021**, 43, 818–830. [Google Scholar] [CrossRef] - Knapp, N.; Fischer, R.; Huth, A. Linking lidar and forest modeling to assess biomass estimation across scales and disturbance states. Remote Sens. Environ.
**2018**, 205, 199–209. [Google Scholar] [CrossRef] - Zhang, G.; Ganguly, S.; Nemani, R.R.; White, M.A.; Milesi, C.; Hashimoto, H.; Wang, W.; Saatchi, S.; Yu, Y.; Myneni, R.B. Estimation of forest aboveground biomass in California using canopy height and leaf area index estimated from satellite data. Remote Sens. Environ.
**2014**, 151, 44–56. [Google Scholar] [CrossRef] - Gao, Y.; Liang, Z.; Wang, B.; Wu, Y.; Shiyu, L. UAV and satellite remote sensing images based aboveground biomass inversion in the meadows of Lake Shengjin. J. Lake Sci.
**2019**, 31, 517–528. [Google Scholar] - Yan, E.; Lin, H.; Wang, G.; Chen, Z. Estimation of Hunan forest carbon density based on spectral mixture analysis of MODIS data. J. Appl. Ecol.
**2015**, 26, 3433–3442. [Google Scholar] [CrossRef] - Kamiyama, M.; Taguchi, A. Color Conversion Formula with Saturation Correction from HSI Color Space to RGB Color Space. IEICE Trans. Fundam. Electron. Commun. Comput. Sci.
**2021**, E104, 1000–1005. [Google Scholar] [CrossRef] - Hunt, R.W.G. The Reproduction of Colour; John Wiley & Sons: Hoboken, NJ, USA, 2004. [Google Scholar]

**Figure 1.**Overview of the study area. (

**a**) Location of study area; (

**b**) Digital orthophoto model (DOM) and distribution of sample sites; (

**c**) Digital elevation model (DEM). (

**d**,

**e**) are ground photographs of the sample plots and trees.

**Figure 2.**Schematic diagram of data collection for (

**a**) UAV equipped with five-lens oblique camera and (

**b**) GCP.

**Figure 4.**Point-like or spot-like abnormal areas after 3D reconstruction. (

**a**) is the DOM data of a sample plots. (

**b**) is the DSM data of a sample plots and the red circle is the point or block-like invalid data.

**Figure 7.**Scatterplots of measured and predicted IT-AGB. Ai, Bi and Ci are models with spatial resolutions of 0.05 m, 0.1 m, 0.2 m, 0.5 m and 1 m, respectively; (

**A1**–

**A5**) are models of experiment A; (

**B1**–

**B5**) are models of experiment B; (

**C1**–

**C5**) are models of experiment C. The red line in the figure is the line of y = x.

**Figure 9.**Schematic diagram of normalization of individual tree point cloud. (The short lines of different colors at the bottom of the left figure represent different DEM values, and the length is the pixel size of the corresponding DEM. The right figure is the top view of the left figure, each solid line grid is 1 m, and different colors represent different DEM values).

Sensors and Flight Parameters | JHP QX3MINI |
---|---|

Sensor size (mm × mm) | 23.5 × 15.6 |

Heading overlap (%) | 80 |

Side overlap (%) | 70 |

Horizontal speed (m·s^{−1}) | 4~8 |

Flight altitude (m) | 100 |

Exposure interval (s) | 0.8~4.5 |

Focal length (mm) | 35 |

Scanning angle (°) | 45 |

Single camera pixel numbers (million) | 42 |

Sensors and Flight Parameters | Parameter Values |
---|---|

Wavelength (nm) | 1550 |

Beam divergence angle (mrad) | 0.5 |

Spot diameter (cm) | 45 |

Pulse repetition rate (kHz) | 360 |

Pulse emission frequency (Hz) | 112 |

Flight altitude (m) | 900 |

Flight speed (m/s) | 55 |

Type | Variable | Formula | Describe |
---|---|---|---|

RGB space | ${{\displaystyle Max}}_{i}$ | —— | Maximum intensity |

${{\displaystyle Min}}_{i}$ | —— | Minimum intensity | |

${{\displaystyle Mean}}_{i}$ | —— | Mean intensity | |

${{\displaystyle sum}}_{i}$ | ${{\displaystyle sum}}_{i}={\displaystyle \sum _{i=1}^{n}(0.2\times {{\displaystyle R}}_{i}+0.72\times {{\displaystyle G}}_{i}+0.07\times {{\displaystyle B}}_{i})}$ | Total intensity | |

${{\displaystyle sum}}_{sqrt\_i}$ | ${{\displaystyle sum}}_{sqrt\_i}=\sqrt{sumi}$ | Square root of intensity value | |

HSI space | ${{\displaystyle I}}_{MAX}$ | —— | Maximum intensity |

${{\displaystyle I}}_{MIN}$ | —— | Minimum intensity | |

${{\displaystyle I}}_{MEAN}$ | —— | Mean intensity | |

${{\displaystyle I}}_{SUM}$ | —— | Total intensity | |

${{\displaystyle I}}_{SUM\_SQRT}$ | —— | Square root of intensity value | |

Hight variables | H_{aad} | $Haad=\frac{{\displaystyle {\sum}_{i=1}^{n}(|Zi-\overline{Z}|)}}{n}$ | Mean absolute deviation, ${Z}_{i}$ is the height of the ith point in each unit, $\overline{Z}$ is the average height of all points, n is the total number of points in each statistical unit. |

H_{AIQ} | —— | Cumulative height percentile interquartile spacing | |

H_{kurtosis} | ${H}_{kurtosis}=\frac{{\scriptscriptstyle \frac{1}{n}}{\displaystyle {\sum}_{i=1}^{n}{{\displaystyle (Zi-\overline{Z})}}^{4}}}{{{\displaystyle ({\scriptscriptstyle \frac{1}{n}}{\displaystyle {\sum}_{i=1}^{n}{{\displaystyle (Zi-\overline{Z})}}^{2})}}}^{2}}-3$ | Height kurtosis | |

H_{cv} | $Hcv=\frac{{{\displaystyle Z}}_{std}}{{{\displaystyle Z}}_{mean}}\times 100\%$ | Variation coefficient of Z value of all points in a statistical unit, Z_{std} and Z_{mean} are the standard deviation of the height values of all points and the average height of all points in each statistical unit. | |

H_{redio} | $Hredio=\frac{Zmean-Z\mathrm{min}}{Z\mathrm{max}-Z\mathrm{min}}$ | Canopy fluctuation rate, Z_{min}, Z_{max}, Z_{mean} are the minimum height, maximum height and average height of all points in each statistical unit | |

H_{stddev} | ${{\displaystyle H}}_{stddev}=\sqrt{\frac{{\displaystyle {\sum}_{i=1}^{n}{{\displaystyle (Zi-\overline{Z})}}^{2}}}{n}}$ | Standard deviation of Z values of all points in a statistical unit | |

H_{variance} | ${{\displaystyle H}}_{\mathrm{var}iance}=\frac{{\displaystyle {\sum}_{i=1}^{n}{{\displaystyle (Zi-\overline{Z})}}^{2}}}{n}$ | Variance of Z values of all points in a statistical unit | |

H_{skewness} | ${{\displaystyle H}}_{skewness}=\frac{\frac{1}{n}{\displaystyle {\sum}_{i=1}^{n}{{\displaystyle (Zi-\overline{Z})}}^{3}}}{{{\displaystyle (\frac{1}{n-1}{\displaystyle {\sum}_{i=1}^{n}{{\displaystyle (Zi-\overline{Z})}}^{2})}}}^{\frac{3}{2}}}$ | Height skewness | |

H_{1…99th} | —— | 1…99% cumulative height percentile | |

H_{max, min, mean, median} | —— | The maximum, minimum, average and median of the point cloud after normalization | |

P_{1st,…,99th} | —— | 75%, 95% high percentile | |

density variables | D_{0,…,9} | —— | The proportion of height points greater than 30%, 50%, 70%, and 90% to all points |

Vegetation index | CIVE | $CIVE=0.44R-0.88G+0.398+18.79$ | Color index of vegetation [53] |

ExG | $ExG=2G-R-B$ | Excess green index [54] | |

ExGR | $ExGR=ExG-1.4R-G$ | Excess green minus excess red index [55] | |

GLA | $GLA=\frac{2\times G-R}{2\times G+R+B}$ | Green leaf algorithm [56] | |

NGRDI | $NGRDI=\frac{G-R}{G+R}$ | Normalized green–red difference index [56] | |

VEG | $VEG=\frac{G}{{{\displaystyle R}}^{a}{{\displaystyle B}}^{(1-a)}},a=0.67$ | Vegetation index [57] | |

COM | $\begin{array}{l}COM=0.25ExG+0.3ExGR\\ +0.33CIVE+0.12VEG\end{array}$ | Combination index [58] |

Sample Number | Number of Trees | Sample Characteristics | DBH (cm) | Tree Height (m) | AGB (kg) |
---|---|---|---|---|---|

1 | 51 | Maximum | 36.90 | 24.50 | 448.85 |

minimum | 10.10 | 9.20 | 21.85 | ||

Mean | 20.90 | 18.53 | 153.23 | ||

Standard deviation | 5.10 | 2.68 | 81.18 | ||

2 | 40 | Maximum | 31.50 | 23.60 | 376.41 |

minimum | 12.50 | 15.80 | 51.88 | ||

Mean | 21.72 | 20.16 | 175.35 | ||

Standard deviation | 5.11 | 2.12 | 84.21 | ||

3 | 28 | Maximum | 33.90 | 25.80 | 386.73 |

minimum | 18.00 | 14.10 | 98.07 | ||

Mean | 25.75 | 19.84 | 227.85 | ||

Standard deviation | 4.11 | 3.25 | 79.23 | ||

4 | 30 | Maximum | 34.80 | 26.10 | 465.611 |

minimum | 5.90 | 4.50 | 19.94 | ||

Mean | 18.55 | 14.19 | 153.89 | ||

Standard deviation | 8.59 | 6.11 | 125.87 | ||

Total | 149 | Maximum | 36.90 | 26.10 | 465.61 |

minimum | 5.90 | 4.50 | 19.94 | ||

Mean | 21.53 | 19.39 | 170.51 | ||

Standard deviation | 6.04 | 3.97 | 92.73 |

Model | Model Equation | R^{2} | RMSECV (kg) | rRMSECV | |
---|---|---|---|---|---|

0.05 m | A | $\begin{array}{l}Y=-508.187+241.946\times GLA+2535.144\times {{\displaystyle D}}_{5}\\ -24.819\times {{\displaystyle H}}_{30\mathrm{th}}+51.368\times {{\displaystyle H}}_{\mathrm{max}}\end{array}$ | 0.44 | 71.05 | 0.40 |

B | $\begin{array}{l}Y=-561.402+720.154\times GLA+2335.538\times {{\displaystyle D}}_{5}\\ -24.972\times {{\displaystyle H}}_{30\mathrm{th}}+51.822\times {{\displaystyle H}}_{\mathrm{max}}+0.090\times {{\displaystyle I}}_{SUM\_SQRT}\end{array}$ | 0.57 | 62.72 | 0.35 | |

C | $\begin{array}{l}Y=-505.267+404.231\times GLA+2057.729\times {{\displaystyle D}}_{5}\\ -21.371\times {{\displaystyle H}}_{30\mathrm{th}}+47.405\times {{\displaystyle H}}_{\mathrm{max}}+0.088\times {{\displaystyle sum}}_{sqrt\_i}\end{array}$ | 0.59 | 61.41 | 0.34 | |

0.1 m | A | $\begin{array}{l}Y=-454.307+600.580\times GLA+2497.540\times {{\displaystyle D}}_{5}\\ -48.036\times {{\displaystyle H}}_{30\mathrm{th}}+70.815\times {{\displaystyle H}}_{\mathrm{max}}\end{array}$ | 0.49 | 67.00 | 0.37 |

B | $\begin{array}{l}Y=-592.935+1219.055\times GLA+2700.810\times {{\displaystyle D}}_{5}\\ -50.306\times {{\displaystyle H}}_{30\mathrm{th}}+75.790\times {{\displaystyle H}}_{\mathrm{max}}+0.111\times {{\displaystyle I}}_{SUM\_SQRT}\end{array}$ | 0.70 ** | 52.74 | 0.30 | |

C | $\begin{array}{l}Y=-556.908+819.871\times GLA+2579.594\times {{\displaystyle D}}_{5}\\ -44.905\times {{\displaystyle H}}_{30\mathrm{th}}+70.367\times {{\displaystyle H}}_{\mathrm{max}}+0.109\times {{\displaystyle sum}}_{sqrt\_i}\end{array}$ | 0.72 ** | 50.10 | 0.28 | |

0.2 m | A | $\begin{array}{l}Y=-350.353+1081.184\times GLA+1891.010\times {{\displaystyle D}}_{5}\\ -49.732\times {{\displaystyle H}}_{30\mathrm{th}}+67.582\times {{\displaystyle H}}_{\mathrm{max}}\end{array}$ | 0.50 | 66.78 | 0.38 |

B | $\begin{array}{l}Y=-529.350+1809.895\times GLA+1919.730\times {{\displaystyle D}}_{5}\\ -48.844\times {{\displaystyle H}}_{30\mathrm{th}}+72.289\times {{\displaystyle H}}_{\mathrm{max}}+0.125\times {{\displaystyle I}}_{SUM\_SQRT}\end{array}$ | 0.75 ** | 48.52 | 0.27 | |

C | $\begin{array}{l}Y=-503.178+1314.658\times GLA+1695.792\times {{\displaystyle D}}_{5}\\ -40.214\times {{\displaystyle H}}_{30\mathrm{th}}+65.158\times {{\displaystyle H}}_{\mathrm{max}}+0.125\times {{\displaystyle sum}}_{sqrt\_i}\end{array}$ | 0.79 ** | 44.77 | 0.25 | |

0.5 m | A | $\begin{array}{l}Y=-326.319+1048.792\times GLA+770.335\times {{\displaystyle D}}_{5}\\ -35.794\times {{\displaystyle H}}_{30\mathrm{th}}+55.757\times {{\displaystyle H}}_{\mathrm{max}}\end{array}$ | 0.44 | 70.39 | 0.40 |

B | $\begin{array}{l}Y=-512.761+1761.952\times GLA+755.589\times {{\displaystyle D}}_{5}\\ -33.174\times {{\displaystyle H}}_{30\mathrm{th}}+59.510\times {{\displaystyle H}}_{\mathrm{max}}+0.124\times {{\displaystyle I}}_{SUM\_SQRT}\end{array}$ | 0.69 ** | 53.99 | 0.30 | |

C | $\begin{array}{l}Y=-498.311+1279.555\times GLA+1678.740\times {{\displaystyle D}}_{5}\\ -25.901\times {{\displaystyle H}}_{30\mathrm{th}}+53.839\times {{\displaystyle H}}_{\mathrm{max}}+0.128\times {{\displaystyle sum}}_{sqrt\_i}\end{array}$ | 0.74 ** | 49.45 | 0.28 | |

1 m | A | $\begin{array}{l}Y=-273.155+553.142\times GLA+184.608\times {{\displaystyle D}}_{5}\\ -17.770\times {{\displaystyle H}}_{30\mathrm{th}}+39.409\times {{\displaystyle H}}_{\mathrm{max}}\end{array}$ | 0.34 | 82.24 | 0.45 |

B | $\begin{array}{l}Y=-441.875+1227.558\times GLA+211.609\times {{\displaystyle D}}_{5}\\ -13.633\times {{\displaystyle H}}_{30\mathrm{th}}+40.918\times {{\displaystyle H}}_{\mathrm{max}}+0.114\times {{\displaystyle I}}_{SUM\_SQRT}\end{array}$ | 0.52 | 71.21 | 0.39 | |

C | $\begin{array}{l}Y=-439.854+829.448\times GLA+201.563\times {{\displaystyle D}}_{5}\\ -7.326\times {{\displaystyle H}}_{30\mathrm{th}}+36.344\times {{\displaystyle H}}_{\mathrm{max}}+0.124\times {{\displaystyle sum}}_{sqrt\_i}\end{array}$ | 0.58 | 66.74 | 0.36 |

Model | Model Equation | R^{2} | RMSECV (kg) | rRMSECV |
---|---|---|---|---|

East | $\begin{array}{l}Y=-186.079+884.733\times GLA+1346.804\times D5\\ +38.489\times {H}_{\mathrm{max}}-29.545\times {H}_{30\mathrm{th}}+0.116\times su{m}_{sqrt\_i}\end{array}$ | 0.29 | 80.79 | 0.45 |

North | $\begin{array}{l}Y=-373.301+880.425\times GLA+1898.861\times D5\\ +47.219\times {H}_{\mathrm{max}}-34.777\times {H}_{30\mathrm{th}}+0.271\times su{m}_{sqrt\_i}\end{array}$ | 0.42 | 73.11 | 0.41 |

South | $\begin{array}{l}Y=-365.544+712.880\times GLA+1217.608\times {D}_{5}\\ +51.278\times {H}_{\mathrm{max}}-30.879\times {H}_{30\mathrm{th}}+0.109\times su{m}_{sqrt\_i}\end{array}$ | 0.58 ** | 61.72 | 0.35 |

West | $\begin{array}{l}Y=-468.532+724.398\times GLA+1875.551\times D5\\ +56.608\times {H}_{\mathrm{max}}-35.762\times {H}_{30\mathrm{th}}+0.183\times su{m}_{sqrt\_i}\end{array}$ | 0.60 ** | 60.17 | 0.33 |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 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 (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Lei, L.; Chai, G.; Wang, Y.; Jia, X.; Yin, T.; Zhang, X.
Estimating Individual Tree Above-Ground Biomass of Chinese Fir Plantation: Exploring the Combination of Multi-Dimensional Features from UAV Oblique Photos. *Remote Sens.* **2022**, *14*, 504.
https://doi.org/10.3390/rs14030504

**AMA Style**

Lei L, Chai G, Wang Y, Jia X, Yin T, Zhang X.
Estimating Individual Tree Above-Ground Biomass of Chinese Fir Plantation: Exploring the Combination of Multi-Dimensional Features from UAV Oblique Photos. *Remote Sensing*. 2022; 14(3):504.
https://doi.org/10.3390/rs14030504

**Chicago/Turabian Style**

Lei, Lingting, Guoqi Chai, Yueting Wang, Xiang Jia, Tian Yin, and Xiaoli Zhang.
2022. "Estimating Individual Tree Above-Ground Biomass of Chinese Fir Plantation: Exploring the Combination of Multi-Dimensional Features from UAV Oblique Photos" *Remote Sensing* 14, no. 3: 504.
https://doi.org/10.3390/rs14030504