suyashnk / 3d-point-cloud-using-the-stereo-images

# StereoDataset.ipynb - Colaboratory

## Overview
StereoDataset.ipynb is a Jupyter Notebook designed for stereo image processing using OpenCV and NumPy. The notebook covers the following key functionalities:

1. Mounting Google Drive to access data.
2. Loading stereo calibration parameters.
3. Reading stereo images from a specified dataset.
4. Performing stereo rectification using the StereoBM algorithm.
5. Generating Disparity Map and Depth Map.
6. Saving Disparity Map and Depth Map as image files.
7. Creating a 3D Point Cloud from the Disparity Map.
8. Visualizing and saving the 3D Point Cloud.

## Instructions
Before running the notebook, ensure you have the required data in thespecified Google Drive path and update the calibration_data dictionary with your stereo calibration parameters.

### Steps
1. Mount Google Drive by uncommenting `!mount /content/drive` (comment out if data is already accessible).
2. Define the output directory for saving results (`output_dir` variable).
3. Load stereo calibration parameters manually in the `calibration_data` dictionary.
4. Load stereo images from the specified file paths.
5. Run stereo rectification using the StereoBM algorithm.
6. Save the Disparity Map and Depth Map as image files.
7. Calculate the Q matrix for reprojectImageTo3D.
8. Generate a 3D Point Cloud and visualize it.

Note: Update the file paths and calibration parameters according to your dataset.

## Warning
A runtime warning may occur during the computation of the Depth Map due to potential divide-by-zero encounters. Handle such warnings appropriately.

---

This GitHub README provides a comprehensive overview of the StereoDataset.ipynb notebook, guiding users through its functionalities and offering clear instructions for setup and execution.