知识卡片

普通读书笔记卡 · 9023 · 生成模型:GPT-6

完整相机矩阵按比例定义

理解等级:中级

对于透视模型,p∼M P_W,M=K[R|−RT] 为3×4矩阵,最终像素由前两项除第三项。M 的全局非零尺度不可观察,所以一般射影相机有11自由度。书中K早期为3×4、后期为3×3,使用前须核对矩阵维数,不能直接混乘。

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

      The concatenation of the extrinsic and intrinsic transformation matrices yields the desired transformation matrix, which will transform world coordinates to rendered pixel coordinates: (39.20) Converting back to heterogeneous from homogeneous coordinates, the 2D pixel indices of a 3D point [X, Y, Z]T are [x, y]T = [x′/w, y′/w]T. We can make all the matrices in equation (39.20) explicit and write: (39.21) The right-most matrix multiplications

    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:摘录中关于完整相机矩阵按比例定义的定义、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:The concatenation of the extrinsic and intrinsic transformation matrices yields the desired transformation matrix, which will transform world coordinates to rendered pixel coordinates: (39.20) Converting back to heterogeneous from homogeneous coordinates, the 2D pixel indices of a 3D point [X, Y, Z]T are [x, y]T = [x′/w, y′/w]T. We can make all the matrices in equation (39.20) explicit and write: (39.21) The right-most matrix multiplications