nky237/image-panorama-stitching

An implementation of image panorama stitching project using SIFT and RANSAC algorithms.

Image Panorama Stitching

Requirements

• Python 3
• python-opencv-contrib <= 4.2.1.16

Note

SIFT is no longer available in opencv > 3.4.2.16, you should choose older versions of opencv-python and opencv-contrib-python, the issue can be solved by using the following command lines:

python -m pip install opencv-python==3.4.2.16
python -m pip install opencv-contrib-python==3.4.2.16

Or just execute:

python -m pip install -r requirements.txt

Usage

1. 2 Images Only (panorama.py)

The application can be run by executing:

python panorama.py "left_image_path" "right_image_path"

Example:

python panorama.py images/left.jpg images/right.jpg

2. Multiple Images (multi_image_pano.py: >= 2 images)

The application can be run by executing:

python multi_image_pano.py --images "folder-containing-images-path" --output "output-file"

Example:

python multi_image_pano.py --images images/3_images --output 3_img_pano.jpg

Results

1. 2-Image Panorama

• Inputs: Left and right images

• Output:

2. 3-Image Panorama

• Inputs: 3 images

• Output:

References

1. https://github.com/linrl3/Image-Stitching-OpenCV
2. https://www.pyimagesearch.com/2018/12/17/image-stitching-with-opencv-and-python/
3. https://medium.com/pylessons/image-stitching-with-opencv-and-python-1ebd9e0a6d78

Nov. 2019

Tran Le Anh