CUDA Image Processing: Grayscale Conversion and Blur

This project demonstrates basic image processing techniques using CUDA and OpenCV. It includes functionality for converting a color image to grayscale and applying a customizable box blur effect.

Features

RGB to Grayscale conversion using CUDA
Customizable Box Blur implementation with CUDA
Multiple blur passes for stronger effect
Efficient use of GPU memory with alternating buffers

Prerequisites

CUDA Toolkit (version 10.0 or later recommended)
OpenCV (version 4.0 or later recommended)
A CUDA-capable GPU
C++ compiler with C++11 support

Building the Project

Ensure you have the CUDA Toolkit and OpenCV installed on your system.
Clone this repository:

git clone https://github.com/yogesh1801/cudaimageprocessor.git

3.Compile the project:

nvcc -std=c++11 main.cu -o image_processor pkg-config --cflags --libs opencv4

# You may use cmake files to directly configure the proect.

Usage

Place your input image in the same directory as the executable and name it img.png.
Run the program: