← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9035.d · 生成模型:GPT-6

单应齐次最小二乘需归一约束

理解等级:高级

求min||Ah||²必须约束||h||=1,排除零解;使用最小奇异值对应右奇异向量再按一致顺序还原H。也可从A^TA的最小特征值向量得到解,但数值稳定性需考虑。H33归一到1只在该项非零且不太小时可用。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      where A is a 2N × 9 matrix, when given N corresponding pairs. The vector h contains all the elements of the matrix H stacked as a column vector of length 9. Note that h only has 8 degrees of freedom as the results do not change for a global scaling of all the values. Therefore we will need at least four correspondences to estimate

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第41章《单应性》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c041.xhtml) - 结论依据:摘录中关于单应齐次最小二乘需归一约束的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:where A is a 2N × 9 matrix, when given N corresponding pairs. The vector h contains all the elements of the matrix H stacked as a column vector of length 9. Note that h only has 8 degrees of freedom as the results do not change for a global scaling of all the values. Therefore we will need at least four correspondences to estimate