Stereo-Depth

This code implements a stereo vision pipeline to calculate the depth of an object in an image.

Dependencies

This code requires the following Python libraries:

• OpenCV
• NumPy
• Matplotlib

Running the Code

1. Clone this repository.
2. Navigate to the project directory in your terminal.
3. Install the required libraries using the following command:

pip install opencv-python numpy matplotlib

4. Run the script using the following command:

python stereoDepth.py

Description

The code performs the following steps:

1. Reads the stereo image pair: It reads the left and right images from a specified file path.
2. Displays the stereo image pair: It displays the left and right images side-by-side using matplotlib.
3. Reads the calibration matrices: It reads the pre-computed camera calibration matrices for the left and right cameras.
4. Computes the disparity map: It computes the disparity map between the left and right images using the SGBM algorithm from OpenCV.
5. Decomposes the projection matrices: It decomposes the projection matrices to obtain the intrinsic and extrinsic camera parameters.
6. Calculates the depth map: It calculates the depth map from the disparity map, camera calibration parameters, and baseline distance between the cameras.
7. Reads the obstacle image: It reads a reference image of the obstacle to be detected.
8. Locates the obstacle in the image: It uses template matching to locate the obstacle in the left image.
9. Calculates the nearest point: It calculates the closest point on the obstacle to the camera based on the depth map and obstacle location.
10. Displays the results: It displays the depth map, bounding box around the detected obstacle, and the depth of the nearest point on the obstacle.

File Descriptions

• stereoDepth.py: This is the main script that executes the stereo vision pipeline.
• files_management.py: This file contains functions to read the images and calibration matrices.
• algo_defs.py: This file contains functions to compute the disparity map, decompose the projection matrices, calculate the depth map, locate the obstacle in the image, and calculate the nearest point.

Results