| 基于极线校正的最小二乘立体匹配 |
| |
| 引用本文： | 李金凤,边继龙.基于极线校正的最小二乘立体匹配[J].微计算机信息,2011(9). |
| |
| 作者姓名： | 李金凤 边继龙 |
| |
| 作者单位： | 牡丹江师范学院计算机科学与技术系;哈尔滨工程大学计算机科学与技术学院; |
| |
| 摘 要： | 本文减少了最小二乘转换参数,通过三个几何转换参数和两个辐射转换参数建立对应关系,采用经极线校正的立体像对,使对应点的搜索在相同扫描行上进行,减小了搜索空间,提高了匹配速度,且把匹配方法嵌入到多尺度空间中以提高匹配速度,通过视差后处理进一步提高匹配精度。采用自适应窗口技术解决由于存在矩阵不可逆情况而导致大量不可匹配点和在地形平坦、灰度变化不明显的区域不匹配或误匹配率高的缺点。试验结果表明了,本算法精度高,匹配率高的优点,有相当的使用价值。
|
| 关 键 词： | 立体匹配 最小二乘法 自适应窗口 多尺度空间 |
Least square stereo matching algorithm based on epipolar constraint |
| |
| LI Jin-feng BIAN Ji-long.Least square stereo matching algorithm based on epipolar constraint[J].Control & Automation,2011(9). |
| |
| Authors: | LI Jin-feng BIAN Ji-long |
| |
| Affiliation: | LI Jin-feng(Dept.of Computer Science and Technology,Mudanjiang Normal University,Mudanjiang 157011,China) BIAN Ji-long(College of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China) |
| |
| Abstract: | Transformation parameters is reduced to three geometrical parameters and two radiometric parameters and search space is reduced by searching correspondences on the same scan line of calibrated stereo images and Least Square method(LSM) is applied in a Multi-scale space? based on Gaussian pyramid decomposition of images in order to improve matching speed.Matching precise is improved by disparity-post processing.The adaptive window is adopted to solve the mismatches owing to nonreversible design matrix and th... |
| |
| Keywords: | stereo matching least square methods adaptive window Multi-scale space |
| 本文献已被 CNKI 等数据库收录！ |
|
