32 lines
1.4 KiB
Markdown
32 lines
1.4 KiB
Markdown
# Binarization
|
|
|
|
## Topic
|
|
|
|
Separation of objects and background in an image with a changing level of light.
|
|
|
|
## Aim
|
|
|
|
Implement an algorithm for block binarization of an image with the choice of
|
|
a separate threshold for each block, evaluate the operation of the algorithm on
|
|
images with varying light levels.
|
|
|
|
## Instructions
|
|
|
|
1. Upload an image with rice grains (rice.tif), find
|
|
the optimal threshold for it using the
|
|
[Otsu](https://en.wikipedia.org/wiki/Otsu%27s_method) method, which will
|
|
separate the images of rice grains from the background, build a histogram of the
|
|
image and mark the threshold found on it
|
|
2. Convert the original image to binary with this threshold, display it on
|
|
the screen
|
|
3. Split the original image into 100 blocks (10 vertically, 10 horizontally),
|
|
find the optimal threshold for each block using the Otsu method, block by block
|
|
convert the image to binary, applying its own threshold for each block
|
|
4. Display the resulting binary image on the screen, compare the results
|
|
5. Repeat steps 1-4 for the image of a geometric figure on the background
|
|
(figure.png)
|
|
6. Make a conclusion about the effect of changes in the illumination level on
|
|
the result of selecting objects against the background during binarization with
|
|
a global threshold, as well as about the possibilities of the block binarization
|
|
method for solving a similar problem, if necessary, provide histograms for
|
|
individual image blocks |