Abstract
Experimental and theoretical measures of flocculation were studied using image analysis. An experimental study of commercial board samples led to the proposal of three descriptive floc features, namely, size, ‘definition’, and contrast. Numerical values were obtained from an ensemble averaged linear auto-correlation function.
In addition a theoretical model of formation was simulated to compare degrees of flocculation. The theoretical structure was created by using a poisson cluster model in conjunction with a coverage model. This led to the superposition of fibrous micro0-flocs whose flock centre radii, R, and fibre content, N, determine the severity of the formation. The variance and p.t.p correlation of the resulting image textures were computed. These measures were found to have a lower limit which is set by the fibrous structure of the flocs.
The findings from the simulation study were then applied in principle to the variance and size information extracted from the board samples to explain their structure. The versatility of programmable image analysis systems was demonstrated for formation measurement.
Download PDF