In this project, I wanted to create a prototype of a household recycling machine. After researching the subject quite a bit, it became clear to me that household recycling is very likely to be automated in the future – it is both more convenient and less error-prone. However, the problem is extremely complex and will require a great team of engineers to complete successfully.
For now, the mechanical design of the bin will remain a secret. Here, I am sharing some code on how the central component of such a bin could work – the material classification system. At the time of the project, no such system was created or open-source. Looking back, RGB computer vision (perhaps, also with the use of depth imaging) is the way to go; however, the tech I used here is also promising for this use.
The backbone of this project is an implementation created by ken1row/libfreenect2_altered.
- Install the necessary software required for libfreenect2 library (https://github.com/OpenKinect/libfreenect2).
- Replace the contents of the /build folder with the contents of this repository.