Flat image from Video
Hi All
Im new here, and currently reading around and learning as much as I can. Im asking to get ideas, help and make connections - im not expecting a compelte answer (although welcomed) Also, if people would be interested in this project, please feel free to contact me
So we have cameras, which travel within pipes, of various sizes. The cameras are looking for defects on the pipe (cracks, bulges and so on)
The video will always have a distance stamp on it (as an overlay) which will tell you the position, as the speed will vary of the travel of the camera within the pipe.
So, for a starter, im wondering whether this set of tools can take this video and effectivly flatten it into a 2d image. The hope is to then wrap this 2d image around a 3d cylinder so you can see the inside of the pipe on the outside, for easier assessment.
So, the obvious question is - this is the right thing for this task, and then any help, opinions or comments are greatly welcomed.
Thanks for reading !
I have queston. Are you putting camera inside pipe or outside?
Inside the pipe
each image from a video IS flat, no ?
it's unclear, what you have and what you attempt to do.
maybe you can add an example image ?
@matyhaty. Are you using snake cable video, similar to surgery? Just taking recorded and then taking to lab to analysis it?
InsideYes, like a snake camera, however its actually on wheels. It crawls within the page, videoing it
so, is this a distortion / undistortion problem ? again, image please.
@matyhaty. Yes, you can do live video that including distance stamp as well as speed, etc. Unfortunately, you can't do live video for opencv for flatten image. You can recording live vdeio and take to lab for analysis for opencv.
So the video is a 'picture' of the inside of a pipe
So, assuming the camera is moving at a constant speed - I would take a section of the image, frame by frame of the video, - undistort it - and it would make a vertical segment of an image.
The next frame would do the same, and so on
Just found this image which is exactly it: https://www.wincan.com/wp-content/upl...
https://docs.opencv.org/master/da/d54...