一种倾斜影像几何纠正的有效方法

Efficient Method of Oblique Image Correction

针对航摄时拍摄角度倾斜于地面而产生的倾斜影像的几何畸变问题,提出了一种基于空间变换的倾斜影像纠正方法。首先根据坐标转换数学模型和新影像的采样间隔,确定出新影像的大小和原影像在新影像中的坐标域;然后在新影像中判断并统计出在坐标域内的像素点,构成坐标域点集;最后推导了转换模型的反变换,建立了关于原影像图像坐标的线性方程组并进行了改进整理,求解方程组的解并转换成像素坐标后,利用最近邻插值获取到坐标域各点的灰度值,以实现影像的纠正。通过Matlab进行了实验,对倾斜影像纠正效果良好,较一般算法相比耗时更少,该研究提出的算法可用于倾斜影像的配准、拼接和建模等处理。

To solve the problem of geometric distortion caused by oblique aerial image,we presented an oblique image correction method based on the spatial transformation.First,the size of the new image and the new image coordinate system of the original image could be obtained according of the mathematical model of coordinate transformation and the sampling frequency of the new image,and then,collections of points were built in the new image coordinate system.Eventually,the inversion transformation model was derived and linear equations about the image coordinate of the original image were established.After solving the equations and transforming the solution into pixels,the digital values of the points were obtained with the nearest neighbour interpolation.Matlab was used in the study,it showed that the results of oblique image correction was good,and processing time was shorter than the traditional method.The method could be used to deal with the oblique image registering,mosaic and modeling.