CV-Road_Segmentation

Overview

This repository is dedicated to the exploration and application of computer vision techniques for two primary tasks: road line extraction and road segmentation. We leverage a combination of techniques to effectively analyze road images and extract useful patterns.

Features

• Region of Interest (ROI) focusing on specific horizon lines.
• Thresholding: Image binarization based on the HSV color scale.
• Morphological Operations: Includes erosion, dilation, opening, and closing for improving the Thresholding results.
• Clustering: Used in segmentation scenarios.
• Canny Edge Detection: Applied for edge detection in images.

Repository Structure

CV-Road_Segmentation │ ├── Clustering_codes/ # Scripts for clustering algorithms
├── Images/ # VArious images for testing and obtained results
├── KITTI_roadDataset/ # KITTI_road dataset images
├── Results/ # Result images for different approaches
├── Videos_results/ # Result videos for different approaches
├── pycache/ # Python cache files
├── carina_dataset/ # Local testing images
├── git_img/ # Images needed to the README.ms file
├── Line_extractionOnVideo.py # Script for video line extraction
├── RoadLine_Extraction.py # Script for extracting road lines
├── Road_Segmentation.py # Script for segmenting roads
├── histogram_plt.py # Script for plotting histograms
├── lower_upper_HSV.py # Script for HSV range adjustment
└── morphological_op.py # Scripts for applying morphological operations

Results

Orginal image:

Thresholding na escala HSV:

Clustering:

Canny Edge Detection:

Line Extraction:

Segmentation: