摘要
扫描笔是一种便携式的文字和图像采集工具,图像拼接技术在扫描笔应用中发挥了重要作用.传统的图像配准方法是基于图像特征点匹配完成图像配准的,但在扫描笔图像拼接中受到采集图像分辨率的影响,该方法不仅耗时长,而且由于特征点的不足,导致难以配准成功.基于该问题,提出了基于相位相关算法对扫描笔图像进行配准,并在此基础上加入了卡尔曼滤波对配准结果进行二次滤波处理.实验结果显示,与传统的图像拼接算法相比,该方法不仅能够保证图像拼接的速度,而且经过卡尔曼滤波后,减少了偏移量中的噪声信息,提高了图像拼接质量.
Abstract
The scanning pen is a portable tool for capturing text and images,where image stitching technology plays a crucial role.Traditional image registration methods are based on matching image feature points to complete the registration.However,these methods are hindered in scanning pen applications due to the influence of image resolution.This method not only takes a long time but also faces challenges in successful registration due to the lack of sufficient feature points.To address this issue,a phase correlation algorithm is proposed for image registration with the scanning pen.Fur-thermore,Kalman filtering is integrated to perform secondary filtering of the registration results.Ex-perimental results demonstrate that compared to traditional image stitching algorithms,this approach not only ensures the speed of image stitching but also reduces noise in offset measurements through Kalman filtering,thereby improving the quality of image stitching.
维普
万方数据