
Can we modify the 360 Bird eye view stitching to just few cameras?
5 views (last 30 days)
Show older comments
Hi,
I would like to modify the 360 degree Bird eye view to just viewing lets say 180 degree view with 3 cameras or even 270 degree view with 5 cameras. The provided example is good enough to run the demo with all the cameras but am looking to see if the above mentioned possibilities exist or the existing function can be modified to work with my expectations!
-Dinesh
0 Comments
Answers (1)
Aravind
on 2 Apr 2025 at 6:35
Yes, it is possible to modify the example at https://www.mathworks.com/help/driving/ug/create-360-birds-eye-view-image.html to use only 5 cameras for a 270-degree view or 3 cameras for a 180-degree view. This is feasible because the operations are performed sequentially with images from each camera. Here is how you can use 5 cameras to obtain a 270-degree view, and you can extend the same logic for 3 cameras to achieve a 180-degree view.
Start by setting the "numCameras" variable to 5. This configuration will select the first 5 cameras, which are numbered in increasing order in a counter-clockwise direction, starting with the top-left camera as 1. With 5 cameras, you achieve an "L" shape coverage on the left side of the car.
Proceed with the example as usual until the "Register and Stitch Bird’s-Eye-View Images" section. Up to this point, you will have obtained the necessary bird’s-eye-view images from each of the 5 cameras. The main change involves registering and stitching these camera images to create a single bird’s-eye-view image. To generate this image, you can use the following code in place of the example's original code:
radius = 15;
% Combine the 5 images from the cameras to generate a single image with 270
% degree view
leftImgs = bevImgs(1:5);
tform = helperRegisterImages(leftImgs, radius);
[leftSideView, Rleft] = helperStitchImages(leftImgs, tform);
figure;
imshow(leftSideView)
This code stitches only the images from the 5 selected cameras, resulting in a 270-degree view instead of a 360-degree one. The result of this can be seen in the image below:

You can apply the above steps to adjust the example for a 270-degree view using any camera arrangement. Just ensure that you correctly adjust the camera indices when loading images in the "Estimate Monocular Camera Extrinsics" section of the example to select the cameras of your interest.
I hope this resolves your query.
0 Comments
See Also
Categories
Find more on MATLAB Support Package for USB Webcams in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!