Open kvnlxm opened 2 years ago
While I am still unsure why this problem exist with the PylonCapture(Pylon)
method, we empirically found out that after a certain amount of time (about 4 minutes with our Pylon settings), the camera feed was finally live. This is good for us at the moment, but I still would like to know why there is such a difference between CamareCapture(Vision)
and PylonCapture(Pylon)
.. Or is there a way to retrieve the Pylon features for the camera also with the CamareCapture(Vision)
method?
Thanks, Kevin
Hi all! Following up on this: Have there been any developments? Currently with the latest Pylon, PylonCapture node, Bonsai version, DLC live install this is still the case and therefore hardly usable...
Best Tobias
To follow up on this: Turns out this indeed was a simple GPU load issue. Binning on the camera solved this.
Hi,
we are currently experiencing an issue with the
PredictPose(DLC)
node for real-time tracking of DLC marker (PreviouslyDetectPose(DLC)
) when combined with thePylonCapture(Pylon)
node. The image output lags about 7 seconds and hence, real time pose-estimation is not possible. In Figure 1PylonCapture(Pylon)
works in real time (withoutPredicPose(DLC)
, while in the setting of Figure 2 it lags. Note that we already tried to remove the conversion node, which does not cause the problem.We also used the
CameraCapture(Vision)
node (orVideoCaptureDevice(Vision)
node), which works in real-time combined withPredictPose(DLC)
, but we cannot retrieve the pylon camera features settings, which is important.Does anyone face the same problem and is there already a work around. My guess is that there is some sort of bug within the
PylonCapture(Pylon)
node and maybe someone from the Bonsai Team can help with that.Cheers, Kevin
Figure 1: PylonCapture works in real-time Figure 2: PylonCapture lags about 7 seconds