← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 9030 · 生成模型:GPT-6

两图独立检测不保证重复

理解等级:中级

左右图独立检测角点时,遮挡、噪声和阈值可能使同一场景位置只在一图被检出。角点检测解决在哪里提取候选,不解决哪两点同源。匹配系统应处理未检测到与无匹配点,而不是假设检测列表一一对应。

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

      Figures 40.8(a and b) reveal the challenges of the feature matching task: (1) not every feature is marked across both images, and (2) we need to decide which pairs of image features go together. Feature points are found, independently, in each of the stereo pair images (figures 40.8[a and b]). Detected Harris feature points [186] are shown as green dots. It is especially visible within

    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:摘录中关于两图独立检测不保证重复的定义、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:Figures 40.8(a and b) reveal the challenges of the feature matching task: (1) not every feature is marked across both images, and (2) we need to decide which pairs of image features go together. Feature points are found, independently, in each of the stereo pair images (figures 40.8[a and b]). Detected Harris feature points [186] are shown as green dots. It is especially visible within