← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9020.a · 生成模型:GPT-6

齐次线性不消除透视除法

理解等级:中级

针孔模型可将 P=[X,Y,Z,1]^T 乘3×4矩阵得到 [fX,fY,Z]^T;去齐次化后仍是 x=fX/Z、y=fY/Z。矩阵形式简化代数,却没有让最终像素对三维欧氏坐标变成线性,也没有从单个像素恢复未知深度。

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

      showing that the matrix K, when multiplied by the homogeneous coordinates of a 3D point, P, renders the homogeneous coordinates of the perspective projection of that 3D point, p. This formulation is one of the biggest benefits of using homogeneous coordinates. It allows writing perspective projection as a linear operator in homogeneous coordinates.

    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:摘录中关于齐次线性不消除透视除法的定义、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:showing that the matrix K, when multiplied by the homogeneous coordinates of a 3D point, P, renders the homogeneous coordinates of the perspective projection of that 3D point, p. This formulation is one of the biggest benefits of using homogeneous coordinates. It allows writing perspective projection as a linear operator in homogeneous coordinates.