ArUco Marker Detection and Real Time Distance Measurement with OpenCV

in this video we will be looking at my code on how opencv works with things like this these are Oracle markers so I'm part of my University Rover team and alongside object detection for the movement of our Rover aruco markers play a huge huge role for the autonomous navigation especially or map based navigation whatever you say and in this video we'll like have a look into how you can actually use the Oracle markers to measure distance like that's the first step of the map based navigation system so let's get started so in this code we are supposed to make 5x5 Oracle markers so they can generate them through the code we can also like achieve the same stuff from like online [Music] no we can run python so this is our Oracle marker 5x5 now we will be generating into this uh generating like 20 of them into this MD folder yeah as you can see we only have this one we are supposed to have 20 as many as we want but we will generate 20 for now so to get the 20 5x5 markers that we want we just comment out this because last time I just wanted you to see how one market is generated so yeah um with that I guess control this I saved the file and run it and as you can see we have 25 X5 markers with starting with 0 to 19 yeah so yeah for the marker generating part that's it and for the detection we can start here a different file um let me get my Oracle markers 5x5 here okay as you can see like these are giving proper IDs for number one number four together in a list number one for together yeah one four and if for this code let me just interrupt thank you nice and for this call give proper IDs sorry for the brightness yeah it works perfectly until I take it too far away and when my camera was detecting this Cube it could do it until 8 9 feet only when the borders were clear so I realized the corners were really important and now for the camera calibration we'll be relying on a 9x6 chessboard design you can find the pattern from opencv's website or any chessboard design would actually work and the goal is we want to train our camera not to take distorted photos and gaining accuracy so let's get into the code and if you want to play with my code on your own compiler make sure you visit my GitHub so yeah and as you can see we have two frames one is the copy frame on the right side and the left side will have the augmentation of course and you can see the photo saving count so yeah all the images are stored into the path we specified in the code as you can see in the folder we have all the pictures the nine by six dimensions are pretty visible now to access the calibration data we will import the OS module and num5 and with the same Z function we will get several errors into one single file uncompressed as you can see the dot npz format which contains all the object points and image points in a more readable Matrix form now let's check if our code works or not and that if we could measure real-time distance with it the distance was accurate accurate until I took a diagonal perspective at least not straight from the camera and the accuracy was dropping to 70 to 75 percent but then again with only 20 photos and the crappy calibration data that resulted in I'm pretty much happy with what we got and if you want if you still want more accuracy you could definitely put in more photos for your reference to get better calibration data better accuracy and you can definitely change the algorithm in your way you could definitely visit my GitHub for that that said I'm pretty much hopeful to make more contents like this so would love it if you listen to Mr Crab because he says you should definitely subscribe to this Channel and you should also listen to Mrs crab because he says you should definitely like this video to help you with the algorithm and you should definitely listen to the couples because you know what if you don't listen to them they might come in your bed at 3am you just don't know so yeah that said goodbye

Channel: SiliconJelly

Views: 4,635

Keywords: opencv, aruco, opencv python, opencv tutorial, aruco detection opencv, aruco marker detection python, aruco detection, aruco markers, aruco marker detection, nasa, mars, perseverance mars rover, mars rover, autonomous navigation, autonomous navigation system, robotics, outdoor autonomous navigation, SiliconJelly, Python Projects, Python Course in Bangla, python course for beginners, Python easiest project, face detection using opencv python, Object Detection, Computer Science, CS50

Id: YOpJrB6bQxo

Length: 5min 42sec (342 seconds)

Published: Fri Feb 24 2023