摘要
针对传统双相机系统,在对可见光图像做畸变矫正后,无法对红外光图像进行矫正的问题,本文提出了一种基于双相机透视系统的矫正靶板图像透视畸变的方法。采用双路CCD相机、靶板、激光照射器和计算机构建激光照射器检测系统,利用标准网格板确定坐标变换矩阵;利用畸变前后的合作目标点确定可见光透视变换矩阵;利用坐标变换矩阵和可见光透视变换矩阵确定红外光透视变换矩阵,进而实现对靶板图像的畸变矫正。实验结果表明:该算法运用坐标变换,矩阵运算,实现了在三维空间中同时矫正双相机透视系统下的可见光图像和红外光图像;不需要测量偏转角度与实时距离,大大降低了实验难度和复杂度;算法可行性高,简单易行,稳定性好。
For traditional dual camera systems,after distortion correction of visible light image,not correcting infrared light image,in this paper,a method of correcting perspective distortion of target image based on dual camera perspective system was presented.The dual CCD camera,target plate,laser irradiator and computer were adopted in laser irradiator detection system.A standard grid plate was used to determine the coordinate transformation matrix;the cooperative target points before and after distortion were used to determine the perspective transmission matrix;the coordinate transformation matrix and visible light perspective transformation matrix were used to determine infrared perspective transformation matrix;and then the distortion correction of the target image was achieved.Experimental results showed that the algorithm used coordinate transformation and matrix operation to achieve simultaneous correction of visible light images and infrared light images in a three-camera perspective system without measuring the deflection angle and real-time distance,which greatly reduced the experimental difficulty and complexity.The algorithm was highly feasible,simple and easy to implement and had good stability.
作者
徐向锴
罗秦
景文博
刘健
XU Xiang-kai;LUO Qin;JING Wen-bo;LIU Jian(School of Optoelectronic Engineering,Changchun University of Science and Technology,Changchun 130022;School of Electronic and Information Engineering,Changchun University of Science and Technology,Changchun 130022)
出处
《长春理工大学学报(自然科学版)》
2019年第5期17-22,共6页
Journal of Changchun University of Science and Technology(Natural Science Edition)
基金
吉林省科技厅项目(20170623004TC
20160204009GX)
关键词
透视畸变
双相机透视系统
合作目标
坐标变换矩阵
Perspective distortion
Dual camera perspective system
Cooperation goals
Coordinate transformation matrix