拼接图像的径向畸变矫正技术

Radial Distortion Correction Technology of Image Mosaic

针对径向畸变引发的拼接图像失真问题,提出一种在图像拼接中引入单参数除法模型的方法。利用Harris算子提取特征点,使用单参数除法模型矫正径向畸变,使用归一化互相关法完成初步匹配,利用随机抽样一致算法提高匹配精度,根据匹配关系完成拼接过程。仿真结果表明:该方法能够快速部署,省去繁重的标定过程,适用于矫正多种类型相机取图产生的径向畸变,实现图像不失真拼接,为先进驾驶辅助、天眼监控等系统的研究工作提供参考。

Aiming at the problem of image mosaic distortion caused by radial distortion,this paper proposes a method which introduces a single parameter division model into image mosaic.In this method,Harris operator is used to extract feature points,and the single parameter division model is used to correct radial distortion.A normalized cross-correlation method is utilized to complete preliminary matching,a random sampling consistent algorithm is used to improve matching accuracy,and the splicing process is completed according to matching relationship.The simulation results show that the method can be deployed quickly,and the heavy calibration process is saved.It is suitable for correcting the radial distortion caused by various types of cameras,which can realize the image mosaic without distortion.It provides a reference for the research of advanced driving assistance system and sky eye monitoring system.