The goal is to blend an image into a background image and extrude the edges in the foreground image to make it visible after matching colors 1:1 with the background image (blending).
Algorithms of interest are Sobel edge detection and weighted sum blending.
Resources
- https://www.meccanismocomplesso.org/en/opencv-python-edge-detection-and-image-gradient-analysis/
- https://docs.opencv.org/3.4/d5/dc4/tutorial_adding_images.html
Sobel edge detection, optimal values:
- ksize (kernel size) = 7
- dx = 1
- dy = 0
Weighted sum blending, optimal values:
- Foreground weight = 55%
- Background weight = 65%
- Scalar added to each sum = 20
| Background Image | Foreground Image |
|---|---|
 |
 |
| Sobel Image | Blended Image (Final Result) |
|---|---|
 |
 |