Water turbidity estimation in water sampled images

This paper tackles the problem of estimating water turbidity by analyzing images. This computer-vision solution avoids to resort to use specific laboratory instruments and, hence facilitates the water characterization in situ. Our contribution consists in designing a whole image processing chain com...

Full description

Saved in:
Bibliographic Details
Published in2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP) pp. 1 - 5
Main Authors Montassar, I., Benazza-Benyahia, A.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.09.2020
Subjects
AdaBoost
deep features
Image segmentation
Mean Shift segmentation
PSPNet
SVM
Online AccessGet full text

Cover

More Information
Summary:This paper tackles the problem of estimating water turbidity by analyzing images. This computer-vision solution avoids to resort to use specific laboratory instruments and, hence facilitates the water characterization in situ. Our contribution consists in designing a whole image processing chain composed of pre-processing, segmentation, feature extraction and classification modules. The second originality of our work relies on comparing two dual approaches for the segmentation and feature extraction: handcrafted and deep neural network based approaches. Finally, the lack of a publicly available dataset has motivated the building of an appropriate dataset. Experimental results indicate satisfactory performances of the proposed approaches.
ISSN:2687-878X
DOI:10.1109/ATSIP49331.2020.9231862