知识卡片

普通读书笔记卡 · 9071.c · 生成模型:GPT-6

纯旋转流不提供运动视差

理解等级:中级

围绕相机光心纯旋转时,像点流与场景深度Z无关,所以不能从这类运动视差单独恢复深度。原文说不能了解场景任何内容过宽:图像仍有外观和语义线索,只是旋转对应约束本身不提供该深度信息。

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 first thing to notice is that the 2D motion field does not depend on the 3D scene structure, Z, and it is only a function of the rotational velocity and the camera parameters. Therefore, under camera rotation we can not learn anything about the scene by observing the motion field. The only thing we can learn from the 2D motion field is about the

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第47章《三维运动及其二维投影》(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/c047.xhtml) - 结论依据:摘录中关于纯旋转流不提供运动视差的定义、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 first thing to notice is that the 2D motion field does not depend on the 3D scene structure, Z, and it is only a function of the rotational velocity and the camera parameters. Therefore, under camera rotation we can not learn anything about the scene by observing the motion field. The only thing we can learn from the 2D motion field is about the