Fig 1.
Overall flow chart of algorithm.
Fig 2.
The WGS84 earth ellipsoid model.
Fig 3.
Schematic diagram of position changes of two images between the previous and latter aerial images photographing by swing-sweep type aerial camera.
Fig 4.
Aeronautical imaging model with soft margin support vector machine.
Fig 5.
Schematic diagram of epipolar constraint of swing-sweep type aerial camera.
Fig 6.
Form of image epipolar constraint on image plane.
Fig 7.
Schematic diagram of optical flow method between aerial camera swing scanning images.
Fig 8.
General diagram of feature matching algorithm.
Fig 9.
Histogram of algorithm run time change.
Fig 10.
Trend chart of matching success rate of sharpness reduction algorithm.
Fig 11.
The clear and the blur original aerial remote sensing images.
Fig 12.
The extraction results of BRISK and FAST algorithm.
Fig 13.
The extraction results of ORB and SURF algorithm.
Fig 14.
Variation curve of pixel speed compensation in X direction calculated by extended L-K optical flow.
Fig 15.
Variation curve of pixel speed compensation in Y direction calculated by extended L-K optical flow.
Fig 16.
Matching results of aerial images with different sharpness by BRISK algorithm.
Fig 17.
Matching results of aerial images with different sharpness by BRIEF algorithm.
Fig 18.
Matching results of aerial images with different sharpness by SURF algorithm.
Fig 19.
Matching results of aerial images with different sharpness by algorithm in this paper.
Table 1.
Compensation amount of pixel moving speed in X direction between different images in swinging strip.
Table 2.
Compensation amount of pixel moving speed in Y direction between different images in swinging strip.
Fig 20.
Matching results of this algorithm for aerial images with different sharpness when the swinging angle is large.
Table 3.
Matching precision of each feature matching algorithm in different image sharpness.
Fig 21.
Histogram of root mean square error of DEM image accuracy versus matching result.