Skip to content

nikhil-nakhate/canny-video-gpu

Repository files navigation

GPU Accelerated Edge Detection in Video

ME-759 High Performance Computation Final Project

By: Sapan Gupta, Nikhil S. Nakhate


Instructions to build the code

Requirements:

  1. Windows 10
  2. Visual studio 2017
  3. CUDA toolkit 9.0.176
  4. CUDA driver 388.16
  5. OpenCV 3.3.1

Environment:

  • OPENCV_DIR = <opencv/build>
  • CUDA_PATH = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.0
  • PATH = %OPENCV_DIR%\x64\vc14\lib;%OPENCV_DIR%\x64\vc14\bin;%PATH%

Instructions to execute the executable:

  • $ cd x64\Release
  • $ hpc759_final_proj_cuda.exe "The Secret Life Of Walter Mitty.mp4"

Executable process description:

This will open two windows for the live original and the processed edge contour feeds. In the edge contour feed, just the edges of the structures and shapes from the original video will get highlighted as white pixels and other areas will be removed as black pixels.

About

Canny Edge detection on video in near real-time

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages