Automated image processing algorithm for 3D OCT images of fouling in spacer-filled membrane filtration channels
Kees Theo Huisman,
Luca Fortunato,
Johannes S. Vrouwenvelder,
Bastiaan Blankert
Affiliations
Kees Theo Huisman
Water Desalination and Reuse Center, Biological and Environmental Science and Engineering Division (BESE), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia; Environmental Science and Engineering Program, Biological and Environmental Science & Engineering (BESE) Division, King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia; College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, Doha, Qatar; Corresponding author.
Luca Fortunato
Water Desalination and Reuse Center, Biological and Environmental Science and Engineering Division (BESE), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia; MANN+HUMMEL Water & Fluid Solutions S.p.A., Italy
Johannes S. Vrouwenvelder
Water Desalination and Reuse Center, Biological and Environmental Science and Engineering Division (BESE), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia; Environmental Science and Engineering Program, Biological and Environmental Science & Engineering (BESE) Division, King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia
Bastiaan Blankert
Water Desalination and Reuse Center, Biological and Environmental Science and Engineering Division (BESE), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia
OCT imaging is an important technique to study fouling in spacer-filled channels of reverse osmosis systems for seawater desalination. However, OCT imaging of membrane filtration channels with feed spacers is challenging because the spacer material can be (partly) transparent, making it difficult to detect and possibly mistaken for fouling, and the longer optical pathway through the spacer material distorts the image below the spacer. This study presents an automated 3D OCT image processing method in MATLAB for visualization and quantification of biofouling in spacer-filled channels. First, a spacer template of arbitrary size and rotation was generated from a CT scan of the feed spacer. Second, background noise and file size were reduced by representing the OCT image with a list of discrete reflectors. Finally, the spacer template was overlayed with the feed spacer in the 3D OCT image, enabling automated visualization of the feed spacer and correction of the distortions. Moreover, the method allows the selection of datasets with the same location relative to the position of the spacer, enabling systematic comparison between datasets and quantitative analysis. • A spacer template of arbitrary size and rotation was generated from a CT scan. • The background noise was removed, and the file size was reduced by representing the OCT dataset with a list of discrete reflectors. • The spacer template was overlayed with the feed spacer in the 3D OCT image.
