基于非线性方程的鱼眼图像畸变矫正方法研究

Distortion correction for fish-eye image based on nonlinear equations

针对鱼眼图像不同区域产生畸变程度各异问题,提出一种非线性方程的鱼眼图像畸变矫正方法,以提高图像矫正精度。首先,利用鱼眼图像自身特性求取其半径与中心位置,并以中心点为圆心将其划分为半径线性递增的同心圆;在此基础上,根据同心圆半径大小为其采用相对应的矫正模型,并通过正交投影法将鱼眼图像映射到半球面上,转化为半球面模型的横向、纵向坐标值,从而投射为以横向与纵向坐标为基础的正方形图像,实现鱼眼图像的精确变换。最后,利用两个实例验证了本文所提出方法的有效性与可行性。

The method of nonlinear equations for fish-eye image distortion correction was proposed to solve the fisheye image distortion in different areas and improve the precision of image rectification. Firstly,the radius and center position were derived by using characteristics of fish-eye image itself. The image could be divided into linear increasing radius of concentric circles. Then,the corresponding correction model was considered by using concentric circle radius value,and the fish-eye images was mapped to hemisphere surface through the orthogonal projection methodology. The horizontal and vertical coordinate values of hemispherical surface model were deduced. So the square image coordinates were described and the fish-eye image precise transformation was realized completely. Finally,two examples verified the feasibility and effectiveness of the proposed results.