I am very new to machine vision but would like to learn about it and wondered what would be the best way to get up to speed. My goal is to add this technology to drones.
What do you want to do? Our system does color tracking the best and AprilTag tracking like stuff.
I have a computer engineering background but completely new to computer vision so I guess I am looking for a good introduction guide to the technology rather than details of OpenMV at this stage. Would it help to first start with OpenCV on a linux platform although OpenMV seems more suited to drones with its lightweight and low power consumption. I would like to explore using this technology to provide stable position control for drones in applications where GPS is unreliable or has failed.
Okay, so, AprilTags would be best for this. Please watch this: April Tags Tracking using the STM32F7 - YouTube
Thanks for the reply and link to the video. It looks like April Tags could be great for precision landing but I really wanted to explore using OpenMV for precision positioning where the camera would use the ground texture. I believe that this is called “optical flow”.
Ah, yeah, then check this video out:
Note that the camera is just giving you one reading. It’s noisy and not that useful by itself unless you filter the output with a kalman filter. In particular, to use such a filter you’ll need to predict the movement of the quadcopter given it’s control inputs along with using another sensor like a gryo to help with additional filtering. Using the OpenMV Cam by itself is not going to provide the best solution.
Thanks for the information. I still need to study this a lot more. Also congratulations on launching the H7 on Kickstarter.