Hey guys,
So I'm in my final year of engineering and for my final project im building an object detection and tracking quadcopter. So a general overview, the idea is that to use a quadcopter's video feed to recognize a pre-decided object and to make the quadcopter move(in what way is not fixed yet) either to avoid it, or to track it.
So to accomplish this I built a very basic quadcopter using the APM 2.6, 1000 kv motors with 10x4.7 props. Im doing this project alone so I fell short on cash just after building the basic quadcopter without the FPV gear and extra CPU to process the images. So I came up with another idea which will just about get the job done. What I plan to do is to use an old smartphone(Samsung galaxy Ace) which weighs about only 110g, install IP webcam on it, and send its video stream to my laptop via a LAN for image processing. I know its not the best option, but this is just a prototype so I just have to demonstrate its operation in a controlled environment and not fly it in different conditions.
I plan on using MATLAB(cause im really fluent in it) for the image processing part and thought of using MAVproxy in order to control the drone.
So here are my questions:
1. Is it practically possible to implement this? I know there will be latency issues but I dont need it to be absolutely quick, I just need to demonstrate the image processing part.
2. This is the main question, how exactly should I go about sending control commands to the APM2.6 from my laptop which will be based on MATLAB's image processing. I have seen people using python scripts to implement this, but I'm not 100% sure weather that would work or not.
Even the slightest amount of help would really be beneficial to me, if someone could just point me in the right direction.
Thanks in advance,
Udai Rathore
Replies