Skip to content

Latest commit

 

History

History
73 lines (45 loc) · 3.22 KB

1-get-an-overview.md

File metadata and controls

73 lines (45 loc) · 3.22 KB
Raw

Step 1: Get an overview

We will as usual start by presenting an overview of the method and the contents of this project.

Holmenkollen dataset

In the "data" directory, we are given a dataset of 110 images taken from a handheld camera inside a helicopter.

For each image, we also have the following data:

  • The intrinsic camera calibration.
  • Navigation data: The pose of the helicopter (which we will call "body") in geographical coordinates (latitude, longitude and altitude).
  • Camera pose measurements: The pose of the camera relative to the helicopter.

We have provided a data reader that we will use to extract these data.

Lab overview

Our job today is to represent and visualise these camera poses in a common coordinate system. We will then project the geographical coordinate of a light pole into the images, so that we can find its pixel position in each image.

The main steps in today's lab are:

  • Represent the geographical body poses in a common local Cartesian coordinate system.
  • Compute the corresponding camera poses in the same coordinate system.
  • Undistort the images using the supplied calibration parameters.
  • Project a geographic world point into the images, using the perspective camera model.

Introduction to the project source files

We have chosen to distribute the code on the following files:

  • main.cpp

    Starts the lab, catches any exceptions and prints their error message on the console.

  • camera_pose_lab.h, camera_pose_lab.cpp

    Runs the main loop. You will implement much of this loop.

  • attitude.h, attitude.cpp

    Represents an attitude (orientation) as a set of x-y-z Euler angles (roll, pitch, yaw). You will be responsible for converting Euler angles to a rotation matrix, represented as a unit quaternion.

  • cartesian_position.h, cartesian_position.cpp

    Represents a position in Cartesian space in meters. You will be responsible for converting this position to an Eigen::Vector3d translation vector.

  • dataset.h, dataset.cpp

    Reads data as a stream of DataElements.

  • geodetic_position.h, geodetic_position.cpp

    Represents a position in geodetic coordinates (latitude, longitude and altitude).

  • intrinsics.h, intrinsics.cpp

    Represents the intrinsic calibration parameters for a camera. You will be responsible for converting these to the camera calibration matrix K and an OpenCV distortion coefficient vector.

  • local_coordinate_system.h, local_coordinate_system.cpp

    Uses the library GeographicLib to represent geographical coordinates in a local Cartesian coordinate system.

  • perspective_camera_model.h, perspective_camera_model.cpp

    Represents a perspective camera. You will be responsible for implementing functionality for projecting 3D points into pixels.

  • viewer_3d.h, viewer_3d.cpp

    Uses the OpenCV module viz3d to visualize results in 3D.

Please continue to the next step to get started!