PART III

三维视觉与几何3D Vision & Geometry

三章、15 节，从像素走到三维世界：相机与投影讲光线怎样变成像素、相机怎样被标定；多视图几何讲两张以上的图像怎样还原位姿与深度；三维表示与配准讲点云、SDF 到 NeRF 与 3D Gaussian Splatting，机器人怎样存放、对齐和渲染它看到的世界。 Three chapters and 15 lessons that go from pixels to the 3D world: cameras and projection covers how rays become pixels and how a camera is calibrated; multi-view geometry covers recovering pose and depth from two or more images; 3D representations and registration covers point clouds, SDFs, NeRF and 3D Gaussian splatting, and how robots store, align and render the world they see.

3 章 · 15 节3 chapters · 15 lessons 前置：Part I（SVD、SE(3)、最小二乘与优化）Prereqs: Part I (SVD, SE(3), least squares & optimization)

1三章之间的关系How the three chapters connect

这三章建立在 Part I（线性代数、概率、李群、优化）和 Part II 的运动学之上，又支撑着后面的状态估计、学习与基础模型。把鼠标放在（或点击）任意一个方块上：琥珀色是它依赖的部分，紫色是用到它的部分；虚线框是前面的章节或后续部分。

These three chapters build on Part I (linear algebra, probability, Lie groups, optimization) and Part II kinematics, and support the state estimation, learning and foundation-model parts that follow. Hover (or tap) any box: amber shows what it builds on, purple shows what builds on it; dashed boxes are earlier chapters or later parts.

MAPPart III 依赖图Part III dependency graph

■ 依赖（先学）builds on (learn first)■ 被用于（后学）used by (learn later)

2各章一览The chapters

★ 标记的是核心节，点击标题即可进入（灰色的是尚未上线的节）。每章都会写出它的"一句话直觉"和在机器人中的主要用途。

★ marks core lessons; click a title to open it (greyed-out lessons are not live yet). Each chapter lists its one-line intuition and its main uses in robotics.

3按目标选路线Pick a route by goal

时间有限的话，可以按你的方向挑一条路线先走通，再回头补全。

Short on time? Pick the route that matches your direction, walk it end to end, then come back and fill in the rest.