Vision-Based Satellite Recognition and Pose Estimation Using Gaussian Process Regression

Table 3

Satellite recognition results on the 1D subset of BUAA-SID 1.5 and comparison with [12, 19]. Results are in bold if the proposed method performs the best and in italics if the second best.

Representation	Method	Accuracy (%)
BI	Homeomorphic manifold [12]	100
	SVM [12]	100
	Kernel method [19]	100
	Our GPR method	100
DT	Homeomorphic manifold [12]	100
	SVM [12]	100
	Kernel method [19]	100
	Our GPR method	100
HU	Homeomorphic manifold [12]	98.9
	SVM [12]	99.6
	Kernel method [19]	94.00
	Our GPR method	99.71
FD	Homeomorphic manifold [12]	95.2
	SVM [12]	98.2
	Kernel method [19]	94.56
	Our GPR method	96
GI	Homeomorphic manifold [12]	100
	SVM [12]	100
	Kernel method [19]	100
	Our GPR method	100
HOG	Homeomorphic manifold [12]	99.9
	SVM [12]	100
	Kernel method [19]	100
	Our GPR method	100