This project aimed to present the development of an initiative on embedded systems , focusing specifically on the implementation of Projection Mapping using an FPGA.
Projection Mapping on an FPGA is a significant technical challenge, involving the use of this chips to calculate and coordinate image or video homographies using a very limited ressources given.
Projection Mapping on an FPGA is a significant technical challenge, involving the use of this chips to calculate and coordinate image or video homographies using a very limited ressources given.
Project specifications :
The main objectives of the project were:
- Implementing rotation and homography algorithms on FPGA.
- Establishing a Projection Mapping solution to display an image on a given surface.
These tasks must be automated using an external camera and a computer, which will perform video processing and send adjustment coefficients via the serial port of our board.
The setup includes:
- Implementing rotation and homography algorithms on FPGA.
- Establishing a Projection Mapping solution to display an image on a given surface.
These tasks must be automated using an external camera and a computer, which will perform video processing and send adjustment coefficients via the serial port of our board.
The setup includes:
- A Zybo Z7 FPGA to handle homography calculations and receive coefficients via the serial port communication .
- A MIPI-CSI connected camera, which will focus on the projected image with buffer management.
- OpenCV library to implement algorithms such as findHomography, which will provide homography matrices on a separate computer .
- Catapult to perform HLS of our C++ code.
- Vivado for FPGA coding.
- Vitis to program the ARM processors to process information from the serial port.
Full details and reference can be found here :