基于小波高精度点匹配的基本矩阵估计

利用2视图进行景物的3维重构中,基本矩阵的估计是一个非常重要的问题。目前无论是线性算法还是非线性算法,在估计基本矩阵时都不太理想。这里采用图像的Harries角点作为端点,经过基于小波变换的子线段匹配方法获得高精度的匹配点对;在此基础上,进一步采用Hartley提出的改进八点算法获得了高精度的基本矩阵。通过真实图像的对比实验证明,该方法能有效地提高景物3维重构中基本矩阵估计精度。

Fundamental Matrix Robust Estimation Based on the Wavelet High Precision Points Match

It's an important problem to estimate the fundamental matrix in two images for 3D construction.Now,it's not perfect to do this both with the liner and non-liner algorithm.The Harris characteristic points were used as extremity to get the high-precision matching points by the way of matching sub-line segment which based on wavelet transformation,and then,normalized eight-point algorithm proposed by Hartley was used to estimate the high-precision fundamental matrix.The true image experiments showed that this ...