This project was developed as part of my Bachelor's degree coursework in Computer Engineering, focusing on Digital Image Processing. The project demonstrates and compares five fundamental edge detection algorithms commonly used in computer vision and image analysis.
Edge detection is a fundamental technique in image processing that identifies boundaries and significant changes in image intensity. This project implements and compares five classical edge detection methods, providing both individual implementations and a comprehensive comparison tool.
- Algorithm: Multi-stage algorithm using Gaussian smoothing, gradient calculation, non-maximum suppression, and double thresholding
- Best for: Clean, well-defined edges with minimal noise
- Advantages: Robust, handles noise well, produces thin edges
- Disadvantages: More computationally intensive
- Algorithm: Uses 3×3 convolution kernels for horizontal and vertical gradients
- Best for: Detecting edges in noisy images
- Advantages: Good noise suppression, directional edge detection
- Disadvantages: May miss some edges, produces thicker edges
- Algorithm: Similar to Sobel but with different kernel weights
- Best for: Edge detection with moderate noise
- Advantages: Good noise handling, isotropic response
- Disadvantages: May produce thicker edges than Canny
- Algorithm: Uses 2×2 convolution kernels
- Best for: Detecting edges at 45-degree angles
- Advantages: Simple, fast, good for diagonal edges
- Disadvantages: Very sensitive to noise, may miss some edges
- Algorithm: Uses second-order derivative operator
- Best for: Detecting fine details and zero-crossings
- Advantages: Detects edges in all directions, good for fine details
- Disadvantages: Very sensitive to noise, may produce double edges
Image-Processing-main/
├── all_together.py # Main comprehensive comparison script
├── canny.py # Individual Canny edge detection
├── sobel.py # Individual Sobel edge detection
├── prewitt.py # Individual Prewitt edge detection
├── roberts.py # Individual Roberts edge detection
├── lap.py # Individual Laplacian edge detection
├── requirements.txt # Python dependencies
├── README.md # This file
├── final.pdf # Project documentation
└── final.pptx # Project presentation
- Python 3.6 or higher
- pip (Python package installer)
-
Clone the repository
git clone https://github.com/shoghli1999/Image-Processing-main.git cd Image-Processing-main -
Install required dependencies
pip install -r requirements.txt
The main script all_together.py provides a complete comparison of all edge detection methods:
python all_together.pyFeatures:
- Interactive image path input
- Side-by-side comparison using matplotlib
- Detailed analysis of each method
- Optional individual OpenCV windows
- Educational explanations and comparisons
You can also run each edge detection method separately:
python canny.py # Canny edge detection
python sobel.py # Sobel edge detection
python prewitt.py # Prewitt edge detection
python roberts.py # Roberts edge detection
python lap.py # Laplacian edge detectionThe comprehensive script provides:
-
Visual Comparison: A 2×3 grid showing:
- Original image
- Canny edge detection result
- Sobel edge detection result
- Prewitt edge detection result
- Roberts edge detection result
- Laplacian edge detection result
-
Detailed Analysis: Printed comparison of each method's characteristics, advantages, and disadvantages
-
Interactive Features: Option to view individual OpenCV windows for detailed inspection
Sobel Kernels:
Vertical: Horizontal:
[-1 -2 -1] [-1 0 1]
[ 0 0 0] [-2 0 2]
[ 1 2 1] [-1 0 1]
Prewitt Kernels:
Vertical: Horizontal:
[-1 -1 -1] [-1 0 1]
[ 0 0 0] [-1 0 1]
[ 1 1 1] [-1 0 1]
Roberts Kernels:
Vertical: Horizontal:
[ 1 0] [ 0 1]
[ 0 -1] [-1 0]
Laplacian Kernel:
[ 0 1 0]
[ 1 -4 1]
[ 0 1 0]
- opencv-python ≥ 4.5.0 - Computer vision library
- numpy ≥ 1.19.0 - Numerical computing library
- matplotlib ≥ 3.3.0 - Plotting and visualization library
This project was developed as part of my Bachelor's degree coursework in Digital Image Processing. The project demonstrates:
- Algorithm Implementation: From-scratch implementation of classical edge detection algorithms
- Comparative Analysis: Systematic comparison of different approaches
- Practical Application: Real-world image processing techniques
- Educational Value: Understanding fundamental computer vision concepts
- Understanding of convolution operations and kernel-based filtering
- Implementation of gradient-based edge detection methods
- Comparison of different edge detection approaches
- Practical experience with OpenCV and image processing libraries
- Analysis of algorithm performance and characteristics
- Educational: All algorithms implemented from scratch for learning purposes
- Comprehensive: Complete comparison of five different methods
- Interactive: User-friendly interface with multiple visualization options
- Well-Documented: Detailed explanations and technical documentation
- Modular: Individual scripts for each method plus comprehensive comparison
- Fixed
cv2.waitkey(0)typo (corrected tocv2.waitKey(0)) in the comprehensive version - Added proper error handling and input validation
- Implemented normalization for better visualization
- Enhanced user interface and documentation
This is an academic project, but suggestions and improvements are welcome! Feel free to:
- Report bugs or issues
- Suggest improvements to the algorithms
- Add new edge detection methods
- Improve documentation
This project is part of academic coursework and is provided for educational purposes.