Figures (6)  Tables (7)

    • Figure 1. 

      Spatial distribution of different complex plots with various forest types. (a) Location of study area in Hubei; (b) plots in subtropical zone, located in Shennongjia, Hubei Province; (c) on-site photos of the 12 plot forest types with different complexity levels; PB and PC represents the planted broad-leaved and planted coniferous forest with a small amount of understory and medium visibility, respectively; NB and NC represent the natural mixed broad-leaved forest, and the natural mixed coniferous forest, with a large amount of understory and poor visibility. Source: http://bzdt.ch.mnr.gov.cn/index.html

    • Figure 2. 

      Schematic diagram of the handheld LiDAR scanning path.

    • Figure 3. 

      Illustration of upper-canopy occlusion in point clouds acquired by HMLS.

    • Figure 4. 

      Flowchart of the proposed method: (a) HMLS point cloud of forest plot with occlusions; (b) stem segmentation and trunk points extraction; (c) stem reconstruction using Kunze's stem curve equation; (d) stand H-D model development; (e) timber and stand volume calculation.

    • Figure 5. 

      Scatter plots of LiDAR-extracted forest attributes vs measured forest attributes across plots: (a) diameter at breast height (DBH), (b) tree height, (c) stem volume, and (d) stand volume.

    • Figure 6. 

      Bar plots of bias errors for LiDAR-extracted vs measured forest attributes across plots: (a) diameter at breast height (DBH), (b) tree height, (c) stem volume, and (d) stand volume.

    • Plot ID Characteristics of plots Number of trees Stem density
      (stems/ha)      		 DBH (cm) Height (m) Slope (°) Canopy cover Stem volume (m3) Stand volume
      (m3/ha)
      1 PB 69 767 15.6 ± 11.1 10.5 ± 5.1 18 0.7 0.27 ± 0.69 141.1
      2 PB 66 733 18.5 ± 12.8 10.9 ± 5.6 16 0.8 0.28 ± 0.50 140.7
      3 PB 56 622 17.1 ± 10.7 10.4 ± 5.3 14 0.5 0.22 ± 0.44 79.2
      4 PC 60 667 26.2 ± 10.2 17.5 ± 6.7 21 0.7 0.66 ± 0.61 429.3
      5 PC 61 678 28.8 ± 6.9 16.2 ± 5.3 23 0.8 0.54 ± 0.39 355.1
      6 PC 56 622 30.8 ± 8.2 21.0 ± 7.3 25 0.6 0.89 ± 0.55 518.3
      7 NB 78 867 19.7 ± 11.9 12.7 ± 5.2 16 0.7 0.36 ± 0.57 206.3
      8 NB 61 678 21.1 ± 14.6 13.9 ± 6.7 17 0.6 0.54 ± 1.13 152.4
      9 NB 48 533 25.4 ± 16.2 11.6 ± 6.1 13 0.5 0.54 ± 1.15 91.5
      10 NC 82 911 25.1 ± 9.7 15.0 ± 7.3 20 0.6 0.49 ± 0.50 410.7
      11 NC 56 622 21.1 ± 8.2 12.8 ± 4.5 22 0.7 0.30 ± 0.39 158.0
      12 NC 88 978 21.9 ± 9.4 12.5 ± 5.3 26 0.8 0.36 ± 0.57 203.1
      Note: plot characteristics are defined by origin (P: planted, N: natural), and forest type (B: broad-leaved, C: coniferous).

      Table 1. 

      Information from forest plot surveys of varying complexity.

    • Plot ID DR (%) CR (%) R2 bias (cm) rbias (%) MAE (cm) rMAE (%) RMSE (cm) rRMSE (%)
      1 91.04 9.52 0.98 0.13 0.90 0.67 4.60 1.31 8.95
      2 91.80 8.20 0.99 0.16 0.99 0.31 1.93 0.66 4.15
      3 94.30 5.66 0.96 −0.09 −0.57 1.10 7.26 1.63 10.71
      4 98.31 1.69 0.97 0.41 1.57 1.05 4.00 1.97 7.53
      5 98.31 1.67 0.97 0.51 1.95 0.94 3.58 1.78 6.81
      6 96.30 3.70 0.97 0.08 0.27 0.67 2.18 1.31 4.27
      7 95.89 4.00 0.98 0.02 0.12 0.54 3.09 1.27 7.22
      8 88.89 10.91 0.99 0.02 0.12 0.11 0.61 0.20 1.11
      9 86.05 14.29 0.99 −0.08 −0.42 0.21 1.08 0.55 2.82
      10 98.68 1.23 0.94 0.67 2.75 1.18 4.81 2.12 8.66
      11 96.15 3.70 0.94 0.29 1.43 1.11 5.58 1.88 9.45
      12 96.20 3.53 0.89 0.76 3.76 1.11 5.47 2.12 10.47
      Note: DR represents detection rate; CR represents commission rate.

      Table 2. 

      The accuracy of DBH in each plot.

    • Plot ID Successful stem curve constructed rate (%) Tree height for stem curve constructed
      R2 bias (m) rbias (%) MAE (m) rMAE (%) RMSE (m) rRMSE (%)
      1 42.62 0.90 −0.20 −2.14 1.20 12.89 1.41 15.09
      2 83.93 0.89 −0.61 −5.77 1.36 12.79 1.63 15.39
      3 84.00 0.89 −0.56 −5.67 1.35 13.72 1.58 16.06
      4 81.03 0.97 −0.17 −0.97 0.92 5.23 1.21 6.92
      5 82.76 0.95 −0.34 −2.05 0.90 5.48 1.23 7.45
      6 88.46 0.97 0.04 0.20 0.95 4.45 1.28 5.97
      7 55.71 0.80 0.24 1.80 1.74 13.07 2.64 19.83
      8 64.58 0.94 −0.44 −3.92 1.15 10.19 1.45 12.86
      9 91.89 0.97 −0.34 −3.29 0.62 6.09 0.96 9.40
      10 74.67 0.97 −0.06 −0.36 0.94 5.88 1.37 8.64
      11 54.00 0.93 −0.62 −4.59 0.88 6.50 1.35 9.94
      12 64.47 0.86 −0.59 −4.91 1.14 9.47 1.54 12.76

      Table 3. 

      The accuracy of tree height from stem curve constructed in plots.

    • Plot ID Tree height from H-D model
      R2 bias (m) rbias (%) MAE (m) rMAE (%) RMSE (m) rRMSE (%)
      1 0.78 −0.71 −6.75 1.33 12.67 2.08 19.80
      2 0.25 −3.28 −24.75 5.25 39.64 8.18 61.72
      3 0.04 −2.07 −16.69 3.60 29.10 5.42 43.82
      4 0.41 −2.18 −12.18 3.27 18.31 3.67 20.53
      5 0.30 0.73 4.78 2.95 19.20 3.73 24.25
      6 0.27 1.17 6.91 2.55 15.00 3.45 20.32
      7 0.68 0.41 3.84 1.46 13.54 1.96 18.19
      8 0.50 −1.80 −12.38 3.29 22.59 3.65 25.10
      9 0.06 −2.87 −19.22 3.92 26.28 4.65 31.17
      10 0.36 0.71 5.02 3.61 25.55 4.91 34.75
      11 0.60 −0.10 −0.80 2.38 18.65 3.02 23.61
      12 0.33 1.43 12.26 2.72 23.32 3.61 30.96

      Table 4. 

      The accuracy of tree height from stand H-D model in plots.

    • Plot ID Stem volume
      R2 bias (m3) rbias (%) MAE (m3) rMAE (%) RMSE (m3) rRMSE (%)
      1 0.96 −0.0240 −11.55 0.03 15.91 0.0978 46.99
      2 0.98 −0.0105 −4.65 0.03 11.44 0.0645 28.52
      3 0.88 −0.0131 −9.19 0.03 19.52 0.0549 38.51
      4 0.96 −0.0067 −1.01 0.06 8.69 0.1205 18.08
      5 0.98 0.0052 0.95 0.03 6.32 0.0579 10.52
      6 0.97 0.0083 0.93 0.06 6.84 0.1020 11.37
      7 0.94 −0.0125 −4.71 0.04 13.44 0.1087 40.99
      8 0.95 −0.0392 −13.70 0.05 17.48 0.1365 47.77
      9 0.97 −0.0115 −5.15 0.02 8.82 0.0460 20.64
      10 0.96 0.0218 4.43 0.06 12.60 0.1011 20.52
      11 0.96 0.0022 0.78 0.04 15.32 0.0748 26.32
      12 0.77 0.0283 11.77 0.05 20.52 0.0970 40.30

      Table 5. 

      The accuracy of stem volume in each plot.

    • Attribute Forest type R2 bias rbias (%) MAE rMAE (%) RMSE rRMSE (%)
      DBH PB 0.98 0.08 0.50 0.68 4.46 1.25 8.21
      PC 0.97 0.29 1.05 0.71 2.55 1.44 5.19
      NB 0.99 0.00 −0.02 0.33 1.86 0.90 5.10
      NC 0.93 0.61 2.81 1.13 5.22 2.06 9.49
      Tree height PB 0.70 −0.77 −7.37 1.64 15.77 2.76 26.47
      PC 0.93 −0.19 −1.04 1.26 6.91 1.85 10.18
      NB 0.83 −0.24 −2.04 1.51 12.79 2.21 18.75
      NC 0.80 0.00 0.03 1.64 12.07 2.53 18.69
      Stem volume PB 0.97 −0.02 −8.34 0.03 14.96 0.08 39.11
      PC 0.97 0.00 0.30 0.05 7.31 0.10 13.88
      NB 0.95 −0.02 −6.47 0.03 13.20 0.08 36.02
      NC 0.95 0.02 5.61 0.05 15.24 0.09 27.09
      Stand volume PB 0.86 −10.04 −8.34 10.04 8.34 10.97 9.12
      PC 0.99 1.28 0.30 4.17 0.96 4.21 0.97
      NB 0.92 −11.77 −7.84 11.77 7.84 13.58 9.05
      NC 0.98 14.44 5.61 14.44 5.61 17.36 6.75
      Units for bias, MAE, and RMSE correspond to the respective attributes: cm for DBH; m for tree height; m3 for stem volume; and m3/ha for stand volume. Other metrics (rbias, rMAE, and rRMSE) are expressed as percentages (%).

      Table 6. 

      The accuracy for forest attributes across four forest types.

    • Method R2 bias (m) rbias (%) MAE (m) rMAE (%) RMSE (m) rRMSE (%)
      LiDAR 360 0.85 −0.40 −3.57 0.61 5.48 0.81 7.33
      Integrated 0.92 −0.19 −1.04 1.26 6.91 1.85 10.18

      Table 7. 

      Accuracy comparison for tree height estimation.