Hi,
I looked through the elbrus.h file and in the Elbrus initialization we pass the ELBRUS_CameraRig which in turn contains the ELBRUS_Camera and the number of cameras in the rig.
/*
* Use this to initialize Elbrus
*
* ELBRUS_TrackerHandle will remember number of cameras
* from rig. Elbrus supports only Mono and Stereo rigs.
*/
ELBRUS_API
ELBRUS_Status ELBRUS_CreateTracker(ELBRUS_TrackerHandle* tracker,
const struct ELBRUS_CameraRig* rig,
const struct ELBRUS_Configuration* cfg);
Also in the code where we pass the images to start the tracking, it is mentioned we need to pass as many images as the number of cameras.
* images - is a pointer to single image in case of mono or
* array of two images in case of stereo.
*
* You should use the same number of images for tracker equal to
* rig->num_cameras in ELBRUS_CreateTracker.
*/
ELBRUS_API
ELBRUS_Status ELBRUS_Track(ELBRUS_TrackerHandle tracker,
const ELBRUS_Image *images,
const ELBRUS_Pose* predicted_pose,
ELBRUS_PoseEstimate* pose_estimate);
So is it possible to use Multiple Cameras in Elbrus SLAM? If yes, how is the multiple camera stream used to build the pose graph?
Hi, I looked through the
elbrus.h
file and in the Elbrus initialization we pass theELBRUS_CameraRig
which in turn contains theELBRUS_Camera
and the number of cameras in the rig.Also in the code where we pass the images to start the tracking, it is mentioned we need to pass as many images as the number of cameras.