• Laser & Optoelectronics Progress
  • Vol. 58, Issue 2, 0210016 (2021)
Jiangang Tu1, Hui Wang1、*, Cheng Xu1, Jinjun Ju1, and Zenghui Shen2
Author Affiliations
  • 1Engineering Equipment Department of Training Base, Army Engineering University, Xuzhou, Jiangsu 221004, China
  • 2Beijing Zhong Ke Zhi Yi Science and Technology Co., LTD, Beijing 100084, China
  • show less
    DOI: 10.3788/LOP202158.0210016 Cite this Article Set citation alerts
    Jiangang Tu, Hui Wang, Cheng Xu, Jinjun Ju, Zenghui Shen. Hyperspectral Image Mosaicking Based on Double-Layer Fusion of Image and Data[J]. Laser & Optoelectronics Progress, 2021, 58(2): 0210016 Copy Citation Text show less
    Framework of image mosaicking algorithm
    Fig. 1. Framework of image mosaicking algorithm
    Schematic of calculation method of weighted sum method
    Fig. 2. Schematic of calculation method of weighted sum method
    Schematic of splitting high and low status data
    Fig. 3. Schematic of splitting high and low status data
    Storage way of BIL
    Fig. 4. Storage way of BIL
    Hyperspectral images to be mosaicked. (a) Image 1; (b) image 2
    Fig. 5. Hyperspectral images to be mosaicked. (a) Image 1; (b) image 2
    Feature points of two hyperspectral images
    Fig. 6. Feature points of two hyperspectral images
    Range of matching feature points
    Fig. 7. Range of matching feature points
    Correspondence between characteristic point pairs
    Fig. 8. Correspondence between characteristic point pairs
    Image extracted in single channel. (a) Image 1; (b) image 2
    Fig. 9. Image extracted in single channel. (a) Image 1; (b) image 2
    Mosaicking results of different images. (a) Low status image; (b) high status image
    Fig. 10. Mosaicking results of different images. (a) Low status image; (b) high status image
    Data of different types. (a) High status data; (b) low status data; (c) hyperspectral image data
    Fig. 11. Data of different types. (a) High status data; (b) low status data; (c) hyperspectral image data
    Stored data
    Fig. 12. Stored data
    Image after mosaicking
    Fig. 13. Image after mosaicking
    Two other hyperspectral images to be mosaicked. (a) Image 3; (b) image 4
    Fig. 14. Two other hyperspectral images to be mosaicked. (a) Image 3; (b) image 4
    Hyperspectral image after mosaicking
    Fig. 15. Hyperspectral image after mosaicking
    Similarity curves of different scene images. (a) Grass land; (b) cement land; (c) road; (d) forest land
    Fig. 16. Similarity curves of different scene images. (a) Grass land; (b) cement land; (c) road; (d) forest land
    FeatureImageSimilarityAccuracy
    Grass land(1)Left image0.99760.9825
    Right image0.9825
    Cement land(1)Left image0.99060.9906
    Right image0.9965
    Road(2)Left image0.89310.8923
    Right image0.8923
    Forest land(2)Left image0.99960.9996
    Right image0.9999
    Table 1. Accuracy before and after data layer image mosaicking
    FeatureImageSimilarityAccuracy
    Grayscale feature(1)Left image0.93220.9275
    Right image0.9275
    Shape feature(1)Left image0.91430.9143
    Right image0.9238
    Texture feature(1)Left image0.92310.9147
    Right image0.9147
    Grayscale feature(2)Left image0.92140.9112
    Right image0.9112
    Shape feature(2)Left image0.93780.9319
    Right image0.9319
    Texture feature(2)Left image0.94740.9287
    Right image0.9287
    Table 2. Accuracy before and after image mosaicking
    Jiangang Tu, Hui Wang, Cheng Xu, Jinjun Ju, Zenghui Shen. Hyperspectral Image Mosaicking Based on Double-Layer Fusion of Image and Data[J]. Laser & Optoelectronics Progress, 2021, 58(2): 0210016
    Download Citation