A poor man's robot arm program written in hacky way.
This is a project about designing a robotic arm capable of grabbing objects wisely by their position and depth. The object detection is computed via computer vision algorithms, and the depth of objects is measured by Intel RealSense depth camera. Furthermore, the framework is implemented in an asynchronous way and written in a very efficient language Rust.
Object Detection | Point Cloud of Depth/Color Image |
---|---|
Auto Grabbing
High Performance
We achieved a fast frame rate by the asynchronous framework implemented by Rust. The robotic arm can work with real-time detection.
Robotic arm: Dobot magician
Controller: Nvidia Jetson Nano
Sensor: Intel RealSense D435
-
Install the Rust toolchain manager, Rustup https://www.rust-lang.org/tools/install
-
Install the nightly Rust
rustup toolchain install nightly
-
Install the following dependent libraries on your OS.
- OpenCV 4.2
- librealsense, please build it with Python binding support.
-
Build & execute
cd arm cargo build --release ./target/release/hacky-arm
-
Directly run
cd arm cargo run --release
Import RabbitMQ and ProtocolBuffer.
We will finally license the source code in some way. Before it happens, the code is under proprieraty license. Please ask for author's consent if you would like to copy or distribute this project.