Abstract
This paper presents a method for predicting positions of color cubes inside a square transparent solid object from images taken at the orthogonal boundary surfaces. The work is developed for use in mapping flow of non-aqueous phase liquids (NAPL) in transparent soils. Transparent soil models have been developed to study the flow of contaminants through porous media, in bench scale tests. Yellow transparent cubes are used to represent NAPL plumes and clear transparent cubes are used as representations of transparent soil in order to definitively validate the algorithm. Color space information is used to relate concentration and image intensity. The new algorithm employs a so-called 3D carving method to iteratively reconstruct a 3D model using images taken at three orthogonal boundaries. The methodology presented in this paper is a fast, relatively accurate, non-intrusive and inexpensive method for quantifying NAPL zones in transparent soil models.
Original language | English (US) |
---|---|
Pages (from-to) | 96-109 |
Number of pages | 14 |
Journal | Computers and Geosciences |
Volume | 78 |
DOIs | |
State | Published - May 1 2015 |
Keywords
- 3D reconstruction
- Colorimetery
- Contaminant transport
- Model studies
- Optical image processing
- Regression analysis
ASJC Scopus subject areas
- Information Systems
- Computers in Earth Sciences