Fish species recognition by shape analysis of images

摘要

Size and species grading of fish (e.g. on board a fishing boat) might in future be done entirely automatically using image analysis. This has been investigated by creating a data bank of shapes from photographs of seven different types of fish on a computer.Three methods of discriminating between the different species of fish are compared and these are: invariant moments, optimisation of the mismatch, and shape descriptors. Each approach was evaluated using discriminant analysis. The methods sorted the fish with a reliability of 73%, 63% and 90% respectively. Further work is required to make a commercial grading system feasible.