摘要
考虑在图象点集含有 Gaussian白噪声的情况下 ,找出平面欧氏变换下 2个点集之间的点匹配。以一定概率选取距离相等的 2对点确定欧氏变换、同时又尽量减少噪声对数值计算的影响。建立了图象噪声传递的模型 。
The point pattern matching problem has many applications in domains such as computer vision and pattern recognition. This paper researches the techniques to eliminate hypotheses from consideration in the alignment method between two planer sets under Euclidean transformation based on probability and numerical stability, thus the computation cost is reduced greatly. To verify the accuracy of the matching and avoid the use of arbitrary thresholds, an noise propagation model is derived to deal with the Gaussian white noise in image.
出处
《武汉理工大学学报》
CAS
CSCD
2003年第6期67-70,共4页
Journal of Wuhan University of Technology
基金
国家自然科学基金资助项目 (6 0 0 730 5 7)
关键词
点模式匹配
欧氏变换
物体识别
Χ^2分布
point pattern matching
Euclidean transformation
object recognition
χ 2 distribution
