基于网格计算的自适应迭代图像配准算法研究

An Adaptive Iterative Image Registration Algorithm Based on Grid-Based Motion Statistics

在基于线阵CCD相机的汽车底盘安全检测中,图像配准质量对于检测结果尤为重要。由于汽车行驶速度的无规律性,汽车底盘图像往往出现大尺度缩放。为提高图像配准质量,本文引入网格移动统计算法,从匹配流程和自动化参数等方面进行了深入研究,提出了自适应迭代的网格移动统计(AI-GMS)算法。实验结果表明,该算法在特征点配准的数量和质量上都有较大提升,纵向完成度能够提高30%以上。

In the vehicle chassis safety inspection based on the linear CCD camera,the research on the image registration algorithm is very important.Due to the irregularity of the speed,the image of the car chassis often appears large-scale zoom.In order to improve the quality of image registration,this paper introduces the GMS(Grid-Based Motion Statistics)algorithm,and conducts in-depth research on the matching process and automation parameters,and proposes adaptive iterative grid motion statistics(AI-GMS)algorithm.Experiments show that the algorithm has greatly improved the number and quality of feature point registration,and the longitudinal completion can be improved by more than 30%.