Bulletin of Surveying and Mapping ›› 2025, Vol. 0 ›› Issue (5): 89-93.doi: 10.13474/j.cnki.11-2246.2025.0515

Previous Articles     Next Articles

The effectiveness of multi-perspective image tilt on the feature matching of ASIFT algorithm

XIE Kun, ZHOU Jie   

  1. Provincial Geomatics Centre of Jiangsu, Nanjing 210013, China
  • Received:2024-09-10 Published:2025-06-05

Abstract: The ASIFT algorithm is an image feature extraction algorithm with complete affine invariance. It handles affine distortions in images by simulating variations in the camera optical axis, demonstrating strong robustness in images matching with large perspective changes.In order to study the influence of different tilt images on the feature matching of ASIFT algorithm in oblique aerial photography,this paper combines the two affine transformation parameters defined in the ASIFT algorithm: absolute tilt t and transition tilt τ. A method is proposed to calculate the absolute tilt of a single image and the tilt of multi-view images in oblique photography. By conducting ASIFT matching experiments under various tilt conditions, the paper analyzes and summarizes the effects of different tilt angles on feature matching,and validate the effectiveness of the multi-perspective image tilt calculation method.

Key words: oblique photogrammetry, ASIFT algorithm, relative tilt, absolute tilt, image matching

CLC Number: