FPGA Implementation of Fog Removal using Anisotropic Diffusion

Hardware-optimized fog removal algorithm for real-time video processing

Project Supervisor: Professor Sudipta Mukhopadhyay (IIT Kharagpur)

Implemented and optimized a fog-removal algorithm on a Xilinx Zedboard FPGA, achieving 200 fps processing speed for real-time video dehazing. The project involved designing the algorithm in synthesizable hardware description language and integrating system-level blocks.

Technical Accomplishments:

  • Framework: Vivado HLS, Vivado IP Integrator
  • Target: Xilinx Zedboard FPGA
  • Algorithm: Anisotropic diffusion-based fog removal
  • Performance: 200 fps in hardware simulations
  • Process: Vivado HLS design → RTL synthesis → system-level integration

This implementation demonstrates the feasibility of real-time video enhancement on embedded hardware platforms, critical for mobile and autonomous systems applications.

Project Report