Image stitching method for the complicated scene of coalmine tunnel based on mismatched elimination with directed line segments

In order to better solve the problem of mismatching of feature points in the complicated scenes of coalmine tunnel and the projection distortion that usually occurs in the process of image stitching, a coalmine tunnel image stitching method based on directional line segment mismatch elimination was proposed. First, the SIFT (Scale Invariant Feature Transform) algorithm was employed for image feature extraction and matching to obtain a rough matching point pairs. Then a directed line segment model of coarse matching point pairs of adjacent images was constructed, and the direction and length attributes of the line segment was used to eliminate mismatched point pairs. After that the directed line segment model and its direction label of the characteristic points were established in the respective images, and then direction matching was performed on the directed line segments corresponding to the adjacent images, and the probability statistical model was used to remove the mismatched point pairs twice to obtain the final fine matching point pair. Finally, an image grid model was established, and the AANAP (Adaptive As-Natural-As-Possible) algorithm was used to align and stitch the images, and the weighted average method was used to fuse the images to complete the image stitching. Feature matching and image stitching experiments were carried out on coalmine tunnel images and four sets of public data sets. Compared with the RANSAC (RANdom SAmple Consensus) algorithm, the algorithm has better real-time performance and higher matching point accuracy. In addition, the registration accuracy of the corresponding coalmine tunnel image stitching is higher, and the panoramic stitching image obtained is more natural. The experimental results show that the proposed algorithm is an image stitching algorithm with high accuracy and good stitching effect for the complicated scene of coalmine tunnel, and it has better robustness and availability.