Image Synthesis Techniques for Feature Extraction

Abstract

The main objective of this dissertation is producing statistical analysis for quality check of membrane filter using Image Processing. Quality of membrane filter is determined by uniformity factor of the polymer beads under test. Membrane filters or "membranes" are micro-porous films with specific pore size ratings. Membranes retain particles and microorganisms that exceed their pore ratings by acting as a physical barrier and capturing such particles on the surface of the membrane. This dissertation aims at producing a system which generates statistical data for quality check of membrane filter using image processing. Uniformity factor of the polymer beads determines the quality of the given membrane filter. Through this work results are presented from a study where the membrane filter is analyzed for its operating functionality using various feature extraction techniques such as Watershed, Active Contouring, Global thresholding, Circular Hough transform, etc. The ultimate goal is to use this information for analyzing the operating efficiency of a membrane filter. The images represent the filter with different no. of beads along with their boundary extracted images. We obtain good results for separate beads along with relatively complex overlapping beads. According to the quality of the membrane filter obtained it can be used for various applications within the industry, science and research. Some of applications are water recovery and recycle, fluid and water purification, capture and recovery of fluid suspended products (biological, minerals etc.), desalination of water, medical and pharmaceutical use, bacteriological examination in air and water and food and dairy industry.