Is there a way to mark the moving points and in the end determine the position of every point? [closed]

asked 2019-04-18 07:41:27 -0500

nikogamulin gravatar image


Suppose there is a camera mounted above the container ship and captures just a certain area of the moving ship. The stream is fed into machine learning model which detects containers of various colours. Once the whole ship is scanned, is there a way to provide the report about the detected container positions, relative to the ship dimensions (or pixels defining the ship), without knowing the speed of the ship while passing the scanner?



edit retag flag offensive reopen merge delete

Closed for the following reason question is off-topic or not relevant by berak
close date 2019-04-20 00:25:55.830668


I've not understand the need of machine learning in this case and also how you can scan an entire ship if the camera is attached on a container. Sorry for misunderstanding. By the way, I think that this is a mix of visual odometry and object detection.

HYPEREGO gravatar imageHYPEREGO ( 2019-04-18 07:58:24 -0500 )edit

This is just a hypotethical example. The main thing I would like to solve is to track arbitrary points and assess their position relative to the moving object.

nikogamulin gravatar imagenikogamulin ( 2019-04-18 08:26:13 -0500 )edit

Python? how many camera? Stitching = Yes or No?

supra56 gravatar imagesupra56 ( 2019-04-18 09:21:39 -0500 )edit

Python is fine. For beginning there's one camera but later I might add more; could apply stitching - doesn't matter - whatever way to get the desired results.

nikogamulin gravatar imagenikogamulin ( 2019-04-18 11:25:31 -0500 )edit