UAV Image Matching Method Integrating SIFT Algorithm and Detection Model Optimization

Sha GAO; Xi-ping YUAN; Shu GAN; Lin HU; Rui BI; Rao-bo LI; Wei-dong LUO

doi:10.3964/j.issn.1000-0593(2022)05-1497-07

Journals >Spectroscopy and Spectral Analysis >Volume 42 >Issue 5 >Page 1497 > Article

Spectroscopy and Spectral Analysis
Vol. 42, Issue 5, 1497 (2022)

UAV Image Matching Method Integrating SIFT Algorithm and Detection Model Optimization

Sha GAO^1;, Xi-ping YUAN^{1; 3;}, Shu GAN^{1; 2; *;}, Lin HU^1;, Rui BI^1;, Rao-bo LI^1;, and Wei-dong LUO^1;

Author Affiliations

1. School of Land and Resources Engineering, Kunming University of Science and Technology, Kunming 650093, China

show less

DOI: 10.3964/j.issn.1000-0593(2022)05-1497-07 Cite this Article

Sha GAO, Xi-ping YUAN, Shu GAN, Lin HU, Rui BI, Rao-bo LI, Wei-dong LUO. UAV Image Matching Method Integrating SIFT Algorithm and Detection Model Optimization[J]. Spectroscopy and Spectral Analysis, 2022, 42(5): 1497 Copy Citation Text

show less

Fig. 1. Flow chart of UAV image matching

Download full size | View in the Article

Fig. 2. Extreme value detection of Gaussian difference function

Download full size | View in the Article

Fig. 3. Experimental image data

Download full size | View in the Article

image pair	SIFT		SURF		BRISK		ORB		KAZE
image pair	Feature points	Times /ms	Feature points	Times /ms	Feature points	Times /ms	Feature points	Times /ms	Feature points	Times /ms
Fig A Fig B	63 570 69 869	2 950.58	29 791 35 970	954.61	33 367 32 242	36 933.4	14 875 14 886	461.37	32 314 39 112	6 634.32

Table 1. Statistics on the number and length of image feature point extraction by different algorithms

View in the Article

image pair	SIFT+		SURF+		BRISK+		ORB+		KAZE+
image pair	Feature points	Times /ms	Feature points	Times /ms	Feature points	Times /ms	Feature points	Times /ms	Feature points	Times /ms
Fig A Fig B	25 040 23 984	899.72	5 380 5 163	109.55	11 585 10 888	392.45	10 827 10 417	200.16	3 520 3 096	106.63

Table 2. Secondary screening of matched point pairs based on Mahalanobis distance and NNDR model

View in the Article

image pair	SIFT+		SURF+		BRISK+		ORB+		KAZE+
image pair	Inliers	matched	Inliers	matched	Inliers	matched	Inliers	matched	Inliers	matched
Fig(A-B)	7 360	8 756	1 258	1 695	2 064	3 000	1 590	2 292	1 171	1 506

Table 3. Statistics related to image feature point matching by different algorithms

View in the Article

image pair	SIFT+		SURF+		BRISK+		ORB+		KAZE+
image pair	Accuracy	RMSE	Accuracy	RMSE	Accuracy	RMSE	Accuracy	RMSE	Accuracy	RMSE
Fig(A-B)	0.841	0.85	0.742	0.91	0.688	1.22	0.694	1.14	0.773	0.86

Table 4. Statistics of correct extraction rate and RMSE rate of image feature points by different algorithms

Download Citation

Save the article for my favorites

Paper Information