OpenCL Image Filters Examples

This repository provides an implementation of various image processing filters using OpenCL and OpenCV. The filters include:

• Gaussian Filter: For denoising images.
• Sharpening Filter: For enhancing image edges.
• Bilateral Filter: For edge-preserving smoothing.
• Mean Filter: For averaging neighboring pixels.
• Median Filter: For removing noise using median values.

Requirements

Software

• OpenCL: Ensure you have the OpenCL runtime installed. For example, the NVIDIA or AMD drivers should include OpenCL support.
• OpenCV: For image loading and saving. Install OpenCV as described below.
• C++ Compiler: A C++ compiler supporting C++11 or later.

Installing Dependencies

OpenCL

1. On Ubuntu:

   sudo apt-get update
   sudo apt-get install opencl-headers ocl-icd-opencl-dev intel-opencl-icd

2. On macOS: OpenCL is included with the system. Ensure you have the latest Xcode command line tools.

3. On Windows: Download and install the appropriate OpenCL runtime from the GPU manufacturer (NVIDIA, AMD, Intel).

OpenCV

1. On Ubuntu:

   sudo apt-get update
   sudo apt-get install libopencv-dev

2. On macOS:

   brew install opencv

3. On Windows: Download and install OpenCV from the official website. Set up the environment variables and include directories as instructed.

Building the Project

1. Clone the Repository:

   git clone https://github.com/ggluo/OpenCL-Image-Filters.git
   cd OpenCL-Image-Filters

2. Compile the Code:

   Ensure you have OpenCL and OpenCV libraries and headers. Compile with:

   g++ -o image_filters main.cpp -lOpenCL `pkg-config --cflags --libs opencv4`

   Adjust include paths and library paths as needed.

Running the Example

1. Prepare Your Image Data:

   Place your input image in the project directory or adjust the file paths in the code. For example, use an image named input.jpg or use the provided org.png.

2. Run the Compiled Binary:

   ./image_filters ./misc/org.png

   The program will process the image using the selected filter and save the output as output.jpg.

Kernel Files

The kernel.cl file contains the OpenCL kernel code for the image processing filters. The kernels include:

Mean Filter

Averages the pixel values in a 3x3 neighborhood.

kernel.setArg(2, 3); // 3x3 kernel

Gaussian Filter

Blurs the image using a Gaussian kernel for noise reduction.

float kernel[3][3] = {
    {1/16.0, 2/16.0, 1/16.0},
    {2/16.0, 4/16.0, 2/16.0},
    {1/16.0, 2/16.0, 1/16.0}
};

Sharpening Filter

Enhances edges by applying a sharpening kernel.

float kernel[3][3] = {
    { 0, -1,  0},
    {-1,  5, -1},
    { 0, -1,  0}
};

Bilateral Filter

Smooths the image while preserving edges using spatial and range weights.

float sigma_s = 2.0f; // Spatial variance
float sigma_r = 0.1f; // Range variance