LiDAR Point Object Primitive Obtaining Based on Multiconstraint Graph Segmentation

Zhenyang Hui; Zhuoxuan Li; Penggen Cheng; Zhaochen Cai; Xianchun Guo

doi:10.3788/LOP231575

Journals >Laser & Optoelectronics Progress >Volume 61 >Issue 10 >Page 1037001 > Article

Laser & Optoelectronics Progress
Vol. 61, Issue 10, 1037001 (2024)

LiDAR Point Object Primitive Obtaining Based on Multiconstraint Graph Segmentation

Zhenyang Hui^1、2, Zhuoxuan Li^1、2, Penggen Cheng^1、2、*, Zhaochen Cai^1、2, and Xianchun Guo^1、2

Author Affiliations

¹School of Surveying, Mapping and Spatial Information Engineering, East China University of Technology, Nanchang 330013, Jiangxi , China

²Key Laboratory of Mine Environmental Monitoring and Improving around Poyang Lake, Ministry of Natural Resources, East China University of Technology, Nanchang 330013, Jiangxi , China

show less

DOI: 10.3788/LOP231575 Cite this Article Set citation alerts

Zhenyang Hui, Zhuoxuan Li, Penggen Cheng, Zhaochen Cai, Xianchun Guo. LiDAR Point Object Primitive Obtaining Based on Multiconstraint Graph Segmentation[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1037001 Copy Citation Text

show less

Fig. 1. Flowchart of multiconstraint graph segmentation

Download full size

Angle of normal vectors. (a) Sketch map of normal vector angle of the neighboring building points; (b) sketch map of normal vector angle of the neighboring vegetation points

Fig. 2. Angle of normal vectors. (a) Sketch map of normal vector angle of the neighboring building points; (b) sketch map of normal vector angle of the neighboring vegetation points

Download full size

Fig. 3. Graph segmentation based on multi-constraints. (a) Result of graph segmentation based on multi-constraints; (b) enlarged version of area I; (c) enlarged version of area II

Download full size

Fig. 4. Study areas. (a) Area1; (b) Area2; (c) Area3; (d) Area4; (e) Area5

Download full size

Fig. 5. Comparison of the point cloud segmentation results processed by different methods. (a) Proposed method; (b) DBSCAN; (c) spectral clustering method; (d) referenced segmentation result

Download full size

Fig. 6. Comparison of average accuracy of point cloud segmentation results

Download full size

Fig. 7. Graph segmentation based on multi-constraint results with different

ς

. (a)

ς = 1 °

; (b)

ς = 10 °

; (c)

ς = 15 °

; (d) reference segmentation result

Download full size

Area	Method	P /%	R /%	F₁ /%
Area1	DBSCAN	61.36	1.02	2.01
	Spectral clustering	68.19	25.91	37.55
	Proposed method	68.78	84.83	75.97
Area2	DBSCAN	59.17	3.95	7.41
	Spectral clustering	73.31	59.95	65.96
	Proposed method	71.93	99.61	83.54
Area3	DBSCAN	99.20	8.19	15.12
	Spectral clustering	81.61	36.58	50.52
	Proposed method	60.62	93.53	73.56
Area4	DBSCAN	46.02	29.31	35.81
	Spectral clustering	54.41	59.04	56.63
	Proposed method	47.10	64.00	54.30
Area5	DBSCAN	89.20	64.54	74.90
	Spectral clustering	64.64	45.20	59.46
	Proposed method	92.63	22.33	35.98

Table 1. Accuracy comparison of point cloud segmentation

Area	Proposed method	DBSCAN	Spectral clustering
Mean	1.526	52.835	73.503
Area1	1.951	17.303	82.117
Area2	2.121	107.587	102.463
Area3	2.645	130.124	134.064
Area4	0.519	5.006	29.237
Area5	0.393	4.157	19.635

Table 2. Comparison of segmentation time

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information