Quantification of dirt in drinking water bottles using OpenCV

Abstract

The paper presents a contour function-based algorithm to identify, detect and quantify the dirt that is present in
drinking bottles due to prolonged usage despite regular cleaning by brushes. The images are taken using a smartphone-based
camera before and after cleaning with a commercially available household brush. The image processing techniques can identify
the areas of dirt effectively. This information can help designers to make suitable changes in the cleaning brush or to use appropriate
cleaning liquid to eliminate this dirt. The dirt can have a detrimental effect on human health when water is consumed from such
bottles. Also, the dirt is not noticeable in normal usage and hence advanced image processing techniques are used to identify the
same. The algorithm can be converted into a device as a part of future work.