Abstract: The appearance of red parrot fish is an important factor of grading, and the grading results determine its market value. At present, the classification of blood parrot fish mainly relies on artificial, with strong subjectivity and low efficiency. This paper introduced the use of machine vision technology to evaluate the parrot fish appearance by detailed elaboration under the invariable condition of aquaculture water, through Matlab2017a camera image acquisition software control, and through the binarization, threshold, de-noising, finding the boundary, processing the collected image, finally introducing the use of metric function to judge whether the fish boundary of closed curve is tend to be rounded so as to complete blood parrot fish shape evaluation. The experimental results showed that this method can automatically complete the classification of blood parrot fish and solve the problem of low efficiency and large error existing in traditional methods.

Key words: machine vision, image processing, red parrot fish, appearance evaluation