kinect/codes/Azure-Kinect-Sensor-SDK/examples/transformation
davoudn 5986417fda reorganising folders. 2024-03-06 21:35:53 +03:30
..
CMakeLists.txt reorganising folders. 2024-03-06 21:35:53 +03:30
README.md reorganising folders. 2024-03-06 21:35:53 +03:30
main.cpp reorganising folders. 2024-03-06 21:35:53 +03:30
transformation_helpers.cpp reorganising folders. 2024-03-06 21:35:53 +03:30
transformation_helpers.h reorganising folders. 2024-03-06 21:35:53 +03:30

README.md

Azure Kinect Transformation Example

Introduction

The Azure Kinect Transformation Example shows the transformations between a color image, a depth image, and a point cloud. It has two modes: the capture mode and the playback mode.

The Azure Kinect Transformation Example demonstrates the use of the Azure Kinect transformation functions k4a_transformation_color_image_to_depth_camera(), k4a_transformation_depth_image_to_color_camera() and k4a_transformation_depth_image_to_point_cloud().

Capture Mode

The capture mode in the example uses the device to record one capture (depth map + color image) and writes two color point cloud files (color_to_depth.ply and depth_to_color.ply) to the directory specified by the user. The directory provided by the user needs to exist. The color point cloud files can be viewed using Meshlab.

The first point cloud file (color_to_depth.ply) is generated by warping the color image into the geometry of the depth camera using k4a_transformation_color_image_to_depth_camera(). The resulting transformed color image provides a corresponding BGRA color reading for every pixel of the depth map.

The second point cloud file (depth_to_color.ply) is generated by warping the depth map into the geometry of the color camera using k4a_transformation_depth_image_to_color_camera(). The resulting transformed depth map provides a corresponding depth reading for every pixel of the color image.

Playback Mode

The playback mode extracts a single frame in a recording into a point cloud. The recording should be captured using K4ARecorder and should contain at least a color channel and a depth channel, along with calibration information. The point cloud is generated by warping the depth map into the geometry of the color camera and is stored in the .ply format. Note that the color channel needs to be in MJPEG format.

Usage Info

transformation_example capture <output_directory> [device_id]
transformation_example playback <filename.mkv> [timestamp (ms)] [output_file]

Examples:

transformation_example capture c:\temp\
transformation_example capture output 1

transformation_example playback output.mkv
transformation_example playback output.mkv 2000
transformation_example playback output.mkv 2000 c:\data\out.ply