知识卡片

普通读书笔记卡 · 9050.b · 生成模型:GPT-6

正交SfM与透视SfM不同

理解等级:中级

Tomasi–Kanade经典方法以正交投影简化从轨迹恢复结构,后来方法处理透视条件。投影模型决定约束和歧义,不能将正交推导直接用于近距离强透视场景。模型近似是否足够,需要用实际重投影与相机条件检查。

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

      One of the first algorithms for structure from motion for rigid bodies was introduced by Tomasi and Kanade [470]. For simplicity, their approach assumed parallel projection (orthographic camera) instead of perspective projection. The algorithm took as input a video sequence captured by a moving camera looking at a static scene. The video was first processed by extracting keypoints and tracking them over time. The input

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第44章《多视图几何与运动恢复结构》(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/c044.xhtml) - 结论依据:摘录中关于正交SfM与透视SfM不同的定义、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:One of the first algorithms for structure from motion for rigid bodies was introduced by Tomasi and Kanade [470]. For simplicity, their approach assumed parallel projection (orthographic camera) instead of perspective projection. The algorithm took as input a video sequence captured by a moving camera looking at a static scene. The video was first processed by extracting keypoints and tracking them over time. The input