Separation of Single Wood Branches and Leaves Based on Corrected TLS Intensity Data

Yuan Sun; Xiuyun Lin; Jinxin Xiong; Guojing Ren

doi:10.3788/CJL202148.0104001

Chinese Journal of Lasers, Volume. 48, Issue 1, 0104001(2021)

Separation of Single Wood Branches and Leaves Based on Corrected TLS Intensity Data

Yuan Sun^1,2, Xiuyun Lin¹, Jinxin Xiong^3、*, and Guojing Ren¹

Author Affiliations

¹College of Forestry, Nanjing Forestry University, Nanjing, Jiangsu 210037, China

²Co-Innovation Center for Sustainable Forestry in Southern China, Nanjing Forestry University, Nanjing, Jiangsu 210037, China

³NanjingLiuhe Pingshan Forest Farm Development Co., Ltd., Nanjing, Jiangsu 211500, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(14)

Fig. 1. Diagram of incidence angle effect converted to height effect

Download full size

Fig. 2. Schematic of scanning experiment. (a) Laboratory calibration experiment; (b) outdoor calibration experiment

Download full size

Fig. 3. Schematic of eliminating distance error between materials

Download full size

Fig. 4. Relationship between number of polynomials N₂ and N₃ and residual module. (a) Multi-distance residual module; (b) multi-height residual module

Download full size

Fig. 5. Comparison before and after correction of each material indoor. (a) Original intensity data; (b) corrected intensity data

Download full size

Fig. 6. Distribution of outdoor standing trees intensity data before and after correction. (a) Ginkgo; (b) soapberry

Download full size

Table 1. White paper height and distance correction model

View table

Table 1. White paper height and distance correction model

Heightmodel forwhitepaper	S /m		2	3	4	5	6	7
	I_S=p₁·S³+p₂·S²+p₃·S+p₄	p₁	0.038	0.038	0.021	0.013	0.010	0.006
		p₂	-0.110	-0.136	-0.083	-0.053	-0.032	-0.022
		p₃	-0.152	-0.040	-0.011	-0.005	-0.015	-0.019
		p₄	0.806	0.747	0.655	0.671	0.692	0.739
		R	0.028	0.019	0.025	0.007	0.004	0.005
Distancemodel forwhitepaper	H/m		0	0.5	1	1.5	2	2.5
	I_S=p₁·H²+p₂·H+p₃	p₁	0.017	0.014	0.004	0.001	0.002	-0.001
		p₂	-0.163	-0.125	-0.007	0.032	0.044	0.071
		p₃	1.064	0.929	0.572	0.394	0.281	0.189
		R	0.028	0.049	0.017	0.026	0.030	0.032

Table 2. Selection of residual modules and polynomial dimensions of all materials

View table

Table 2. Selection of residual modules and polynomial dimensions of all materials

Species		Residual module						Power
Species		1	2	3	4	5	6	Power
Soapberry	Height model for back of leaves	0.021	0.024	0.01	0.013	0.004	0.013	3
	Distance model for back of leaves	0.011	0.026	0.028	0.028	0.036	0.017	2
	Height model for front of leaves	0.047	0.028	0.015	0.015	0.017	0.016	2
	Distance model for front of leaves	0.014	0.022	0.024	0.035	0.038	0.028	2
	Height model for branch	0.048	0.033	0.027	0.024	0.011	0.017	2
	Distance model for branch	0.009	0.018	0.019	0.006	0.046	0.064	3
Ginkgo	Height model for back of leaves	0.012	0.006	0.004	0.005	0.006	0.016	2
	Distance model for back of leaves	0.176	0.115	0.111	0.188	0.163	0.206	3
	Height model for front of leaves	0.004	0.007	0.005	0.011	0.012	0.010	4
	Distance model for front of leaves	0.010	0.010	0.013	0.010	0.006	0.009	2
	Height model for branch	0.032	0.041	0.019	0.019	0.017	0.008	2
	Distance model for branch	0.020	0.029	0.033	0.038	0.020	0.043	2

Table 3. Range before and after correction of each materials

View table

Table 3. Range before and after correction of each materials

Species	Row data			Corrected data
Species	Max	Min	Range	Max	Min	Range
Back of soapberry leaves	0.432	0.182	0.250	0.207	0.177	0.031
Front of soapberry leaves	0.425	0.200	0.224	0.193	0.140	0.053
Soapberry branch	0.505	0.188	0.317	0.222	0.175	0.046
White paper	0.800	0.334	0.466	0.578	0.510	0.068
Ginkgo branch	0.506	0.215	0.291	0.280	0.240	0.040
Back of ginkgo leaves	0.267	0.153	0.115	0.067	0.062	0.005
Front of ginkgo leaves	0.295	0.163	0.132	0.090	0.084	0.007

Table 4. Distribution characteristics of point reflection intensity

View table

Table 4. Distribution characteristics of point reflection intensity

Species	Max			Intersect
Species	Row data	Leave	Branch	Row-leave	Row-branch	Leave-branch
Ginkgo	0.175	0.035	0.065	0.095	0.120	0.055
	0.165	0.035	0.050	0.080	0.100	0.050
	0.160	0.040	0.080	0.090	0.130	0.065
	0.200	0.060	0.150	0.110	0.175	0.095
	0.190	0.050	0.080	0.105	0.130	0.070
Soapberry	0.090	0.040	0.055	0.065	0.075	0.095
	0.090	0.050	0.065	0.075	0.080	0.120
	0.090	0.030	0.055	0.135	0.075	0.095
	0.090	0.030	0.050	0.045	0.075	0.100
	0.115	0.065	0.055	0.075	0.090	0.115

Table 5. Threshold for leaf and branch separation
View table
Table 5. Threshold for leaf and branch separation
Model Ginkgo Soapberry
Row data 0.16 0.09
Leaves 0.04 0.05
Branch 0.07 0.07

Table 6. Classification accuracy of branches and leaves of each model unit: %
View table
Table 6. Classification accuracy of branches and leaves of each model unit: %
Model Ginkgo Soapberry
Row data Leaves model Branch model Row data Leaves model Branch model
Leaves 35.032 73.224 49.461 32.414 72.812 52.330
Branch 37.406 75.780 46.084 28.103 55.091 46.488

Table 7. Classification accuracy of species and models unit: %
View table
Table 7. Classification accuracy of species and models unit: %
Species Accuracy of row data Accuracy of leaves model Accuracy of branch model
Ginkgo 85.645 91.504 89.071
Soapberry 83.333 84.323 83.592

Table 8. Classification error of leaf model
View table
Table 8. Classification error of leaf model
Species MAE RMSE RE RRSE
Ginkgo 0.0388 0.1431 7.9065 28.883
Soapberry 0.1397 0.2685 28.7238 54.451

Tools

Get Citation

Copy Citation Text

Yuan Sun, Xiuyun Lin, Jinxin Xiong, Guojing Ren. Separation of Single Wood Branches and Leaves Based on Corrected TLS Intensity Data[J]. Chinese Journal of Lasers, 2021, 48(1): 0104001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites