Cropping Algorithm Based on Crop Image Sequence
<p><span style= · International Journal of Multimedia Computing · 2020
Image stitching technology involves computer graphics, pattern recognition, image processing, and other aspects. It is an interdisciplinary topic and has always been a research hotspot in image processing and computer vision. It is widely used in video understanding and analysis, target detection and tracking, image processing and analysis. Therefore, the research on image stitching technology is very meaningful. This paper mainly studies the stitching algorithm based on crop image sequences. In this paper, according to the transformation matrix calculated by the original SURF algorithm, the rotation angle of the image is 66.89°, and the rotation angle obtained by the algorithm is 65.69 °. The difference between the two is about 1°, and the error is only 1.8%. Accurate results are also relatively high in accuracy. It can also be seen from the comparison of the overall transformation matrix that the two algorithms are very close in the obtained matrix results. In addition to the relatively accurate rotation angle, the error of the translation amount is also about one pixel, achieving high accuracy. Sex. It can also be seen from this experiment that the algorithm in this paper still maintains good results when the rotation angle reaches more than sixty degrees. Experimental results show that the algorithm in this paper can significantly improve the visual effect of the fused image and obtain ideal image output.