← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9024.d · 生成模型:GPT-6

DLT必须排除零解

理解等级:高级

有噪声时求 min ||Am||² 需加 ||m||=1,否则零向量就是无意义最优解。解为 A 最小奇异值对应右奇异向量,等价于 A^TA 最小特征值的特征向量。称“最小特征向量”应明确按特征值选择;归一化坐标和SVD有助数值稳定。

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

      As the measurements of the points in correspondence will be noisy, it is generally useful to use more than six correspondences. Therefore, the solution will not be 0. The solution is the vector that minimizes the residual ∥Am∥2, which can be obtained as the smallest eigenvector of the matrix ATA. Once we estimate m we can rearrange the terms into the projection matrix M. This

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第39章《相机建模与标定》(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/c039.xhtml) - 结论依据:摘录中关于DLT必须排除零解的定义、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:As the measurements of the points in correspondence will be noisy, it is generally useful to use more than six correspondences. Therefore, the solution will not be 0. The solution is the vector that minimizes the residual ∥Am∥2, which can be obtained as the smallest eigenvector of the matrix ATA. Once we estimate m we can rearrange the terms into the projection matrix M. This