知识卡片

普通读书笔记卡 · 9028.a · 生成模型:GPT-6

视差反深度关系有几何前提

理解等级:中级

相同有效焦距、平行光轴、水平基线及已对齐的主点约定下,视差δ=x_L−x_R满足Z=f_x B/δ。f_x与δ都用像素,B与Z同长度单位;图像坐标方向不同会改变符号。一般未校正双目不能直接用水平差当反深度。

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

      We will start with a very simple setting where both cameras are identical (same intrinsic parameters) and one is translated horizontally with respect to the other so that their optical axes are parallel, as shown in figure 40.5. We also will assume calibrated cameras. The geometry shown in figure 40.5 reveals the essence of depth estimation from the two cameras. Let's assume that there is

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第40章《立体视觉》(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/c040.xhtml) - 结论依据:摘录中关于视差反深度关系有几何前提的定义、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:We will start with a very simple setting where both cameras are identical (same intrinsic parameters) and one is translated horizontally with respect to the other so that their optical axes are parallel, as shown in figure 40.5. We also will assume calibrated cameras. The geometry shown in figure 40.5 reveals the essence of depth estimation from the two cameras. Let's assume that there is