High-speed and accurate image matching lays the foundation for the successful detection of printed product defects. This paper proposes an improved Harris corner detection method, which is tested with a stability evaluation criterion to prove the superiority of the operator in industrial environments. An affine transformation model is established to approximate the geometric transformation of the neighborhood of the corresponding feature points, and the model parameters are calculated using a deterministic annealing method to avoid time-consuming exhaustive search. The RANSAC method is used to robustly estimate the basic matrix and homography matrix, and epipolar geometry constraints and homography constraints are established to eliminate mismatched pairs in the initial matching. The algorithm has a fast processing speed and has been successfully applied to the online detection system for printed product defects. Keywords: feature points; image matching; epipolar geometry; homography
You Might Like
Recommended ContentMore
Open source project More
Popular Components
Searched by Users
Just Take a LookMore
Trending Downloads
Trending ArticlesMore