Integration of Electrical Resistance Tomography and Morphological Image Processing for Early Detection of Dairy Fouling

Gihan Kuruppu^*, Mihiraj Karunanayake, Wei Yu, Brent Young

Chemical and Materials Engineering Department, University of Auckland, Auckland, New Zealand

^*Email: kkur847@aucklanduni.ac.nz

ABSTRACT

Early detection of milk fouling in dairy processing remains a critical challenge due to the subtle and progressive nature of deposit formation, which can compromise product quality and process efficiency. This study presents a novel approach towards the early detection of milk fouling by integrating Electrical Resistance Tomography (ERT) with outlier detection. A closed-loop milk circulation system was engineered to simulate fouling conditions by utilizing a transparent, nonconductive vertical pipe equipped with a sixteen electrode inline ERT sensor. A saline solution with equivalent conductivity was also tested to establish a stable reference for comparison. Raw voltage data were post-processed to generate two dimensional cross-sectional conductivity maps, which were subjected to a statistical image analysis involving construction of a pixel-wise statistical baseline, application of Z-score analysis, and morphological filtering. This framework allows for automated identification of conductivity anomalies attributable to early-stage stage fouling without the use of complex machine-learning models. The proposed method successfully identified spatial and temporal fouling patterns, even in very thin fouling layers, demonstrating its potential effectiveness in early fouling detection.

Keywords: Milk Fouling, Fouling Detection, Electrical Resistance Tomography, Image Processing

Industrial Application: Dairy Industry