A Tracking Drift Correction Method Based on Feature Points Matching
Download as PDF
DOI: 10.23977/icamcs.2018.053
Corresponding Author
Xinming Zhou
ABSTRACT
In order to improve the efficiency and stability of tracking drift correction, a tracking drift correction method based on feature point matching is proposed in this paper. Tracking drift detection is judged by the similarity between real-time image target template and reference image target template. If the similarity degree is low, the tracking drift phenomenon occurs and the current tracking results need to be corrected. The affine transformation parameters of the middle target are calibrated and then matched with the reference template to determine the correction results of the tracking points.
KEYWORDS
Tracking drift, cross-correlation template matching, feature point matching