Detecting droplets in an image

Hi
Is there a way to detect the droplets in the attached image and draw a boundary around them with an acceptable accuracy?
My final purpose is to find the area of each droplet so that I can calculate their equivalent circular droplets with the same area.
Thank you for your help.

1 件のコメント

Sindar
Sindar 2020 年 9 月 28 日
the image segmentation app and/or the edge function might work

サインインしてコメントする。

回答 (1 件)

Raunak Gupta
Raunak Gupta 2020 年 10 月 2 日
編集済み: Raunak Gupta 2020 年 10 月 2 日

0 投票

Hi,
There are two image processing methods that can be useful for doing this. You can use regionprops for finding the connected components (in this case droplets) and its relevant property.
For very close to circular shaped droplets you can detect circle in the image using imfindcircles. An example workflow is explained here.

3 件のコメント

Taher Abbasiasl
Taher Abbasiasl 2020 年 10 月 3 日
Thank You for your answer.
I am familar with these methods and I have used them in my codes. My main problem is the bright spots on the droplets which makes it impossible for me to binarizare the image in a way that the droplets could be white (or black) and the background could be black (or white).
Unfortunately I don't have a chance to take the video without those bright spots on the droplets.
Raunak Gupta
Raunak Gupta 2020 年 10 月 3 日
Hi Taher,
I agree the lighting condition make it difficult to use the image processing techniques as evident in your case. In that case I would suggest using the deep learning techniques related to semantic segmentation if you have good amount of training data. Also you can try the contrast enhancement techniques from improving the contrast in the image inturn helping the use of regionprops.
Taher Abbasiasl
Taher Abbasiasl 2020 年 10 月 3 日
Thank you for your time Raunak.
I have almost used all of these techniques and unfortunately I could't get enhanced enough images so that I could detect the droplets with their diameters.
I guess the only way is to change the way of visualization in the first place.

サインインしてコメントする。

カテゴリ

ヘルプ センター および File ExchangePolymers & Soft Matter についてさらに検索

質問済み:

2020 年 9 月 27 日

コメント済み:

2020 年 10 月 3 日

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by