基于交比不变和间距比不变的线阵相机畸变标定方法

Distortion Calibration Method for Linear Cameras Based on Cross Ratio and Spacing Ratio Invariability

基于线阵相机双目视觉成像技术可有效检测汽车底盘异物,保证人员安全。但相机畸变影响几何量测、三维重建精度。并且由于线阵相机与面阵相机的成像模型不同,造成适用于面阵相机畸变校正算法无法直接运用在线阵相机当中。通过研究面阵相机的畸变成像模型,结合线阵相机的成像特点,推导出线阵相机的畸变成像模型。利用交比不变性和间距比不变性,结合线阵相机的畸变成像模型,建立二元畸变方程组;并利用等间距条形标靶获得相应参数。最后对新方法进行了真实图像校正实验,证明该方法畸变标定简单易行,有效。

Linear camera based binocular stereo vision can effectively detect the foreign matters on the chassis and ensure the safety of person. The camera distortion calibration is one of the important factors affecting the accuracy geometric measurement and 3D reconstruction. Because of the different model for linear camera and area camera, the distortion correction algorithms of area camera and its calibration target cannot be directly used in linear camera. A new approach for distortion calibration is presented, which utilizes the cross ratio invariability of perspective projection, spacing ratio invariability and the imaging model of linear camera to build a unary quadratic equation and uses the equal interval target to obtain the corresponding parameters. Digital imitation experiment and real distorted image correction prove this approach is simple, accurate.