• Laser & Optoelectronics Progress
  • Vol. 58, Issue 24, 2410003 (2021)
Pengnan Liu1, Dongdong Xu2、*, and Chunmeng Bai2
Author Affiliations
  • 1Shandong Gold Mining (Laixi) Co., Ltd., Qingdao, Shandong 266000, China
  • 2School of Information and Control Engineering, China University of Mining and Technology, Xuzhou, Jiangsu 221000, China
  • show less
    DOI: 10.3788/LOP202158.2410003 Cite this Article Set citation alerts
    Pengnan Liu, Dongdong Xu, Chunmeng Bai. Scale-Invariant Feature Transform-Based Heterogeneous Image Registration Method[J]. Laser & Optoelectronics Progress, 2021, 58(24): 2410003 Copy Citation Text show less
    Gaussian pyramid and Gaussian difference pyramid
    Fig. 1. Gaussian pyramid and Gaussian difference pyramid
    Schematic diagram of extreme point detection
    Fig. 2. Schematic diagram of extreme point detection
    Experiment results of edge and point feature extraction. (a) Original image; (b) phase consistency method; (c) Sobel; (d) Canny
    Fig. 3. Experiment results of edge and point feature extraction. (a) Original image; (b) phase consistency method; (c) Sobel; (d) Canny
    Generation of descriptors in SIFT algorithm
    Fig. 4. Generation of descriptors in SIFT algorithm
    2×2+2 sub areas proposed in this paper
    Fig. 5. 2×2+2 sub areas proposed in this paper
    Rotation process of different shapes
    Fig. 6. Rotation process of different shapes
    Experiment results of feature point extraction. (a) Proposed method; (b) SIFT; (c) SAR-SIFT; (d) PRO-SIFT
    Fig. 7. Experiment results of feature point extraction. (a) Proposed method; (b) SIFT; (c) SAR-SIFT; (d) PRO-SIFT
    Feature matching results. (a)(b) Proposed method; (c)(d) SIFT; (e)(f) SAR-SIFT; (g)(h) PRO-SIFT
    Fig. 8. Feature matching results. (a)(b) Proposed method; (c)(d) SIFT; (e)(f) SAR-SIFT; (g)(h) PRO-SIFT
    Type of imageProposed methodSIFTSAR-SIFTPRO-SIFT
    Number of matching pointsTime /sNumber of matching pointsTime /sNumber of matching pointsTime /sNumber of matching pointsTime /s
    Country59910.1115412.52712.9534520.44
    Field107211.0438715.211412.5445531.95
    Forest260426.60136052.803602.612076188.77
    Indoor2272.121362.731412.681893.98
    Mountain17708.2140913.02942.2844120.92
    Old building7294.403299.452163.6169118.09
    Street149111.1048912.451772.4275419.86
    Urban23009.18135418.126854.07124821.22
    Water198112.53109624.104563.38174955.73
    Table 1. Comparison of matching results of 9 kinds of NIR and VIS images
    Pengnan Liu, Dongdong Xu, Chunmeng Bai. Scale-Invariant Feature Transform-Based Heterogeneous Image Registration Method[J]. Laser & Optoelectronics Progress, 2021, 58(24): 2410003
    Download Citation