宾大 计算机视觉课程

介绍

• robotics perception
  
  • coursera上的课程
  • 侧重于相机模型和3D有关的部分.
  • 非常适合入门摄像机模型
  • 该项目还有其他机器人课程，总共6门.

我的笔记

• 图像单应性实现贴图实例
• Dolly zoom滑动变焦
• 对极几何–基本矩阵F
• augmented reality(AR)入门实例

课程目录

week1. gemetry of image formation
1 introduction
2 camera modeling
3 single view geometry
4 more on perspective projection
5 glimpse on vanishing points
6 perspective projection1
7 perspective projection2
8 point-line duality
9 rotations and translations
10 pinhole camera model
11 focal length and dolly zoom effect
12 intrinsic camera parameter
13 3D world to first person transformation
14 how to computer instrinsics from vanishing points
15 camera calibration

week2. projective transformations
1 vanishing points; how to computer camera orientation
2 compute projective transformations
3 projective transformations and vanishing points
4 cross ratios and single view metrology
5 two view soccer metrology

week3. pose estimation
1 visual features
2 singular value decomposition
3 RANSAC: random sample consencous1
4 where am I 1
5 where am I 2
6 Pose from 3D point correspondences: the procrestes probem
7 Pose from projective transformations
8 Pose form point correspondences P3P

week4. multi-view geometry
1 epipolar geometry1
2 epipolar geometry2
3 epipolar geometry3
4 RANSAC: 2
5 nonlinear least squares1
6 nonlinear least squares2
7 nonlinear least squares3
8 optical flow: 2D point correspondences
9 3D velocities from optical flow
10 3Dmotion and structure from multiple views
11 visual odometry
12 bundle adjustment 1
13 bundle adjustment 2
14 bundle adjustment 3