On Thu, Dec 16, 2010 at 1:30 PM, papadakis wrote: > > Hi, > > So we've been also trying to experiment with the vslam package that is > provided (with a lot of excitement.. ;-)). > These are the issues that we had to deal: > > * Synchronization of two individual 1394 cameras. Since they run as two > separate nodes, there is no synchronization. We started with a suboptimal > solution just to have some results, but we would like a suggestion on what > is the best way to achieve this in ROS. We have also experimented with a > bumblebee camera where we don't have this problem, but we don't want to > restrict ourselves in a particular camera. You may want to take a look on this topic: https://code.ros.org/discuss/ros-users/browse_frm/thread/2568/10763 Regarding this issue. Long story short: If your camera synchronization is hardware based, two different processes will just work and you will only need to use either an intermediate node to match the timestamps (see https://github.com/laas/ros-hueblob/raw/master/src/nodes/fake_camera_synchronizer_node.cpp) or to patch stereo_image_proc to use approximate time matching instead of exact one (see https://code.ros.org/trac/ros-pkg/ticket/4217). If you do not have any kind of hardware synchronization, you wil have much more trouble. One way would be to wait for camera1394 to be reimplemented as nodelets. Grabbing should be much more atomic than with two different processes. -- Thomas Moulard http://www.linkedin.com/in/moulard