nullptr-deref / GuardianBot

Little bot's software able to detect bypassing humans and create database from their face credentials

Guardian Bot

What for?

Guardian Bot is a research project aimed to investigate computer vision integration into robotic and semi-automated systems. Hope you will enjoy looking through this project or find something new or useful!

Feedback is very appreciated.

All in all

This project uses OpenCV Net module (especially with Caffee dnn framework) and has Serial port communication support. It provides UI which shows the output of neural network facial recognition and also provides serial port communication interface.

Development notes

For a couple of months further my aim is to rewrite this project to make it able to use different cameras, not only RealSense devices (simply because I'm currently have no access to device I used to have).

Build and run

Finally, support for simple cameras delivered so feel free to experiment and if you find any issues using specific camera, please contact me or create a PR.

Build

Then, the build itself:

• First of all, you should have Conan package manager installed on your machine so the project's dependencies can be built. If you doesn't have it, visit conan's official website and install it.

• If you have Conan installed on your computer, then run the following commands:

$ mkdir build
$ cd build
$ conan install .. --build=missing

P.S. For now it is highly recommended to use Visual Studio 16 toolchain because I did not experiment with other toolchains and currently I'm really caring about Windows and Linux support so I'll definetely test current build system with gcc toolchain (but not right now).

• After all dependencies were built and installed, then run standard CMake configuration sequence: run cmake .. under build directory and cmake --build . after configuration is complete.

Run

To run the application you have to go to directory containing fresh-baked executable and type the following command:

./GuardianBotApp -p <path_to_prototxt> -m <path_to_caffee_file>

• The -p or --prototxt command line argument is used to provide path to .prototxt file for dnn model.
• The -m or --model command line argument is used to provide path to Caffee model file itself.

Both files are placed in the repository's root directory and you can use them as a default configuration. Of course, you can use your own but consequences are unknown to me, it's your field for researches.:)

Congratulations! You've successfully started my little application, feel free to explore and upgrade it.

Important note!!!

Program uses GLSL shaders placed under resources/ directory in the project's root directory and also being installed under runtime output directory while build into folder with the same name. Make sure you're running application from project's root directory or runtime output directory so the program can access shaders code!

Contributions

Now there is only one contributor (it's me, of course) but if you have an idea how to improve this project or to make it a little more serious than it is - feel free to contact me or to create your pull requests, any help is appreciated.