Accuracy Assessment of Object-Oriented Classification Based on Regular Verification Points

Xunqiang Gong; Xinglei Liu; Tieding Lu; Dan Liu

doi:10.3788/LOP57.241102

Laser & Optoelectronics Progress, Volume. 57, Issue 24, 241102(2020)

Accuracy Assessment of Object-Oriented Classification Based on Regular Verification Points

Xunqiang Gong^1,2、*, Xinglei Liu^1,2, Tieding Lu^1,2, and Dan Liu²

¹Fundamental Science on Radioactive Geology and Exploration Technology Laboratory, East China University of Technology, Nanchang, Jiangxi 330013, China

²Faculty of Geomatics, East China University of Technology, Nanchang, Jiangxi 330013, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(7)

Fig. 1. Diagrams of regular verification points and random verification points. (a) Regular verification points; (b) random verification points

Download full size

Fig. 2. Fusion image of GF-2

Download full size

Fig. 3. Creation of regular verification points. (a) N=50; (b) N=100; (c) N=150; (d) N=200; (e) N=250; (f) N=300

Download full size

Fig. 4. Comparison of classification results of three methods. (a) Overall classification accuracy; (b) Kappa coefficient

Download full size

Table 1. SVM classification results under different numbers of verification points

View table

Table 1. SVM classification results under different numbers of verification points

Number of verification points	OA of regular verification points /%	OA of random verification points /%	Kappa coefficient of regular verification points	Kappa coefficient of random verification points
50	65.10	46.30	0.440	0.205
100	77.85	62.42	0.642	0.438
150	87.11	59.06	0.808	0.395
200	87.92	86.36	0.810	0.801
250	80.54	63.76	0.693	0.462
300	83.22	61.74	0.734	0.450

Table 2. CART decision tree classification results under different numbers of verification points

View table

Table 2. CART decision tree classification results under different numbers of verification points

Number of verification points	OA of regular verification points /%	OA of random verification points /%	Kappa coefficient of regular verification points	Kappa coefficient of random verification points
50	86.58	61.07	0.798	0.455
100	85.90	83.22	0.795	0.751
150	86.58	83.90	0.801	0.767
200	91.94	86.58	0.874	0.810
250	88.60	81.20	0.830	0.732
300	87.25	81.81	0.810	0.748

Table 3. KNN classification results under different numbers of verification points

View table

Table 3. KNN classification results under different numbers of verification points

Number of verification points	OA of regular verification points /%	OA of random verification points /%	Kappa coefficient of regular verification points	Kappa coefficient of random verification points
50	81.56	65.60	0.742	0.499
100	80.41	83.84	0.713	0.769
150	81.96	72.00	0.737	0.586
200	94.63	86.58	0.918	0.810
250	82.99	71.20	0.751	0.595
300	83.33	81.81	0.766	0.748

Tools

Get Citation

Copy Citation Text

Xunqiang Gong, Xinglei Liu, Tieding Lu, Dan Liu. Accuracy Assessment of Object-Oriented Classification Based on Regular Verification Points[J]. Laser & Optoelectronics Progress, 2020, 57(24): 241102

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Imaging Systems

Received: Apr. 27, 2020

Accepted: May. 22, 2020

Published Online: Dec. 1, 2020

The Author Email: Gong Xunqiang (xqgong1988@163.com)

DOI:10.3788/LOP57.241102

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology