Demo program to save point cloud and rectified image data. #385
Conversation
|
Thank you! |
|
This demo program was based off rgb_depth_aligned but now I noticed the |
| frameDepth = None | ||
|
|
||
| if cv2.waitKey(1) == ord("q"): | ||
| # save all images |
There was a problem hiding this comment.
maybe we could save images when pressing s as save instead?
| queueEvents = device.getQueueEvents( | ||
| ("rgb", "depth", "rectifiedLeft", "rectifiedRight") | ||
| ) | ||
| for queueName in queueEvents: | ||
| packets = device.getOutputQueue(queueName).tryGetAll() | ||
| if len(packets) > 0: | ||
| latestPacket[queueName] = packets[-1] | ||
|
|
There was a problem hiding this comment.
This logic of getting frames isn't ideal. Maybe jut use get() instead, or maybe add syncing logic to it (based on sequence number)?
| stereo.rectifiedRight.link(xoutRectifRight.input) | ||
|
|
||
| # Connect to device and start pipeline | ||
| with device: |
There was a problem hiding this comment.
since you are using context manager, move with dai.Device(pipeline) as device here (like in all other demos)
|
@marco-paladini yes I agree I would just expand the existing demo to add saving option when pressing eg. |
Moving the luxonis/depthai-python#655 pull request to the experiments repo.
This PR provides users with a debugging program that computes a point cloud and writes all debugging data to disk.
Also added a minimal pcd viewer and a README for the demo program.