Installing OpenCV on RPi is a big job! The tracker code are in the opencv contrib repository, and most of the prebuilt programs don't contain this function. So we had to build it on our own. It's not surprising that it takes you more than 2 hours. This website gives us proper cmake command so we can fortunately compile OpenCV.
this website tell us how to use RPi camera with OpenCV.
Give us a overview of OpenCV trackers, and provide a basic tracking code.
We have not done this, but if you want to accelerate the tracker, this may help.
We found yolo here.
We found NNPACK for darknet here.
We found the basic OpcnCV darknet usage here.
We didn't use this in our project. But if you want to use some other neural network, this may help.
We found how to use python to send video onto a website here.
We have not done this, but if you want to control the car by method other than ssh, this may help.
I encountered the same problem while opening python program. I found the solution here.
We plan to use website to control the car, and flask and python is a simple way to build a website. Unfortunately, our tracking program is written in c++. So we have to find a IPC(inter process communication) to help us. Socket, or UNIX Domain Socket, seems good for us, but we haven't combine it with our program. Maybe you can try.
4. UNIX Domain Socket IPC-第 37 章 socket编程
Makefile error: boost/asio.hpp: No such file or directory (Windows)
使用Boost.Asio編寫通信程序
socketserver — A framework for network servers
C++ OpenCV image sending through socket
If the car can auto start web server and tracking program while power on, it will be more convenient. We have not success yet, but you can try!
raspberry pi (linux pi) 如何設定開機自動執行 sell script 呢?
Raspberry Pi - run program at start-up