How RealSense multicam frames are arranged?
I am exploring Intel's RealSense multicam code here:
https://github.com/IntelRealSense/librealsense/tree/master/examples/multicam
I can see that the frames are nicely arranged inside the window. However, I do not see how these frames are arranged by code, except this line:
app.show(render_frames);
Could someone enlighten me on this? Thanks.
-
It looks to me as though the work necessary for generating the frame for each stream may be occurring in the line preceding the app.show(render_frames) instruction:
render_frames[frame.get_profile().unique_id()] = colorizers[serial].process(frame);
Reading through the multicam tutorial line by line, it seems to me that the program is doing the following:
1. Detecting a unique ID number for each attached camera device.
2. Specific streams were not defined in the header of the script with configuration (cfg) instructions. And so when the pipe is started, it uses a Default configuration and generates streams from that device based on the recommended default configuration for that particular camera model. So in the example program, there are five devices attached (presumably 5 unique IDs), and it generates a set of Default streams for each device that totals 18 streams from those five devices.
The panel generation therefore seems to be an automated process based on how many cameras have been attached and what model of camera they are.
I would assume that the formatting is just a standard 'wraparound' effect, with boxes of the same size placed on screen in sequence and the box at the edge of the screen being pushed down to the start of the next row if there is not room for it on the end of a row in the screen space available.
Please sign in to leave a comment.
Comments
1 comment