Vol. 7, Issue 2, Part A (2025)

Autonomous mobile robots and driverless vehicles are being advanced quickly. And image matching play important role in Simultaneous Localization and Mapping, it is necessary to improve matching accuracy. In this paper, a new affine transformation model-based matching algorithm was proposed for high accuracy, maintaining acceptable operating time. First, strongest feature matching pairs were extracted from general feature detectors. And then, affine transformation model was made from them to identify scale and rotation relationship between input and reference images. Finally, all matched pairs were tested through proposed model and mismatched pairs were removed. The efficiency of the proposed algorithm is illustrated by MATLAB simulations. Especially, this was efficient for pattern contained images.