Abstract:
Current shrimp quality evaluation is a subjective sum of visual, smell and texture characteristics, and the manual determination of count and uniformity ratio of a sample batch. The automation of this process is desirable for a more rapidly and objectively repeatable evaluation. As a first step towards this goal, the count and uniformity ratio of whole, headless, peeled-tail-on, and peeled-tail-off tiger and white shrimp were evaluated by a machine vision procedure. The experimental weight and view area were correlated by three different equations (linear, power, and forced-power curves) to find the best correlation. The pixel intesities as it relates to lightness-darkness values were also correlated to see if different forms and species could be distinguished by this method. Data, equations and statistics for the correlations are presented.