This little program is amazing congratulations. I have noticed that it works best when the calibration is done by severely underexposing the table. I went as low as -3 EV. On the contrary going +1 or +2EV makes the program fail every single frame. Is there a way to set up the output bitrate of the video? In my case I've noticed that the video looses almost 30% of the filesize. At the same time the program is super fast so eighter it is not re-encoding the frames or it is doing it with some very fast preset which one is the case? Eventually will it be possible to change the encoding parameters (like slow-fast profile, NVENC or CPU encoding, bitrate or quality). Does the program support 10bit videos?
This is amazing. Davinci Resolve has a really high powered Lens Distortion function in its fusion tab. It lacks the ability to automatically detect lens characteristics and one must eyeball it. I need exact settings though, as I do compositing, which requires me to undistort video, then redistort later, after adding assets. Would this software you created, be able to cough out properties which can be keyed into Resolve? It seems like a very fast and efficient way of doing things
I have executed the calibration and the video result is aweome. I got intrinsics parameters of my camera, and now i would like to use them to correct frames from a videostream using my camera but the unistorted image using these parameters is not correct like the result obtained during the calibration process. i think i am mistaking about how to use the parameters. so i want to ask you how would i use them
can this run on google drive inside their notebook? if yes, can you do an instructional video how to install in notebook and how to use it ? Thanks so much
Can this tool fiz lens distortion in real time? Like, as soon as the frame is captured by the camera it is corrected and sent to my application post processed?
to be able to find a pattern from an image, the pattern's dimension must match exactly as predefined, which in this code is hard-coded as 10x7, or in reality 11 by 8. Also, the entire pattern must be visible in each image.
I didn't add source codes because I figured they're redundant for most viewers on youtube since it's set to work out of the box. do you need them thou?
@@walkwithme2064 yep. it would be priceless, not only to learn how it is put together, but also to be able to extend it to have a couple of functionalities i would love to have, like a 4 point unwrap and size tool after undistorting a image
ok. I’ll add a branch with source, thou I might need some days ‘cause it needs a bit of organizing & more comments before going public. I’m also in middle of a project at work, so..
@@walkwithme2064 no problem. whenever you can is awsome :) i will fork the repo and add my changes and share back. if you find it usefull i can make a pull request after your revison
@@X3msnake a new branch called 'with_source' is added to the git repository. the two programs are in a form of a visual studio project. also, it only contains an opecv dll file for release build, so if you need a debug build, a opencv dll for debug is needed. let me know if you have anything else that I could help. not really neatly commented nor organized, but shouldn't be too bad to read along. I had to copy the source from the parent directory into build folder just to add it in the same git repos differentiated by branch, so some visual studio settings could be a bit weird. sorry for that..
Dure, this is exactly what I needed for a fisheye I have!
This little program is amazing congratulations. I have noticed that it works best when the calibration is done by severely underexposing the table. I went as low as -3 EV. On the contrary going +1 or +2EV makes the program fail every single frame. Is there a way to set up the output bitrate of the video? In my case I've noticed that the video looses almost 30% of the filesize. At the same time the program is super fast so eighter it is not re-encoding the frames or it is doing it with some very fast preset which one is the case? Eventually will it be possible to change the encoding parameters (like slow-fast profile, NVENC or CPU encoding, bitrate or quality). Does the program support 10bit videos?
This is amazing!. Using OpenCV, can i set the variables in a .txt so from there after the camera corrects video feed in real time?
This is amazing. Davinci Resolve has a really high powered Lens Distortion function in its fusion tab. It lacks the ability to automatically detect lens characteristics and one must eyeball it. I need exact settings though, as I do compositing, which requires me to undistort video, then redistort later, after adding assets. Would this software you created, be able to cough out properties which can be keyed into Resolve? It seems like a very fast and efficient way of doing things
Sadly, I have no knowledge on how to make a plugin for either Davinci or Premiere, but I’d happy to see anyone else to do it with the shared codes.
@@walkwithme2064 An ofx plugin would work in a large amount of softwares including Resolve Fusion. Can you export distortion data?
you are an angel for this thank you so much
This is brilliant and I haven't even tried it yet! Thank you. Liked and subbed!
와 대박입니다 감사합니다. 모니터의 OSD가 한글이네요 ㅎㅎ Thank you so much! Brilliant program.
Man! U R super! Thanks a lot! It's great program!
I have executed the calibration and the video result is aweome. I got intrinsics parameters of my camera, and now i would like to use them to correct frames from a videostream using my camera but the unistorted image using these parameters is not correct like the result obtained during the calibration process. i think i am mistaking about how to use the parameters. so i want to ask you how would i use them
can this run on google drive inside their notebook? if yes, can you do an instructional video how to install in notebook and how to use it ? Thanks so much
I want the video you have taken through your gopro, in my case the calibration is not taking place properly.
can we undistort photos???
Great, you really should do a mac version :)
Can this tool fiz lens distortion in real time? Like, as soon as the frame is captured by the camera it is corrected and sent to my application post processed?
This is brilliant! Thank you.
Amazing! Can we get the source code for it?
sure. the source code is available on github repo. there’s a separate branch with the source codes.
I came, I saw and I subscribed🤘
Nice move🤘 thanks
I'm trying to use your program, but it fails to find any pattern. Would you be able to help me troubleshoot?
to be able to find a pattern from an image, the pattern's dimension must match exactly as predefined, which in this code is hard-coded as 10x7, or in reality 11 by 8. Also, the entire pattern must be visible in each image.
@@walkwithme2064 thank you! I definitely didn't have the entire pattern visible the whole time, and didn't know about the size requirement.
안녕하세요 혹시 동영상이 아니고 이미지도 가능한가여 렌즈에 디스토션이 있어서 찍은 많은 이미지들을 자동으로 보정이 가능한가요 ?
네 가능은 한데, 지금은 동영상만 읽고 쓰는 가정하에 만들어서 사진파일을 쓰려면 프로그램 수정이 좀 필요할 것 같네요
Only binaries? is source code available?
I didn't add source codes because I figured they're redundant for most viewers on youtube since it's set to work out of the box. do you need them thou?
@@walkwithme2064 yep. it would be priceless, not only to learn how it is put together, but also to be able to extend it to have a couple of functionalities i would love to have, like a 4 point unwrap and size tool after undistorting a image
ok. I’ll add a branch with source, thou I might need some days ‘cause it needs a bit of organizing & more comments before going public. I’m also in middle of a project at work, so..
@@walkwithme2064 no problem. whenever you can is awsome :)
i will fork the repo and add my changes and share back. if you find it usefull i can make a pull request after your revison
@@X3msnake a new branch called 'with_source' is added to the git repository. the two programs are in a form of a visual studio project. also, it only contains an opecv dll file for release build, so if you need a debug build, a opencv dll for debug is needed. let me know if you have anything else that I could help.
not really neatly commented nor organized, but shouldn't be too bad to read along.
I had to copy the source from the parent directory into build folder just to add it in the same git repos differentiated by branch, so some visual studio settings could be a bit weird. sorry for that..
Too much effort to do