Bivalve shellfish grading method and device

Abstract

The invention provides a bivalve shellfish grading method and device. The method comprises the following steps: acquiring a shellfish image of to-be-graded bivalve shellfish; inputting the shellfish image into a trained shellfish grading model to obtain a shell contour in the shellfish image, wherein the trained shellfish grading model is obtained by training an improved convolutional neural network through a shellfish image sample, and a feature matrix in the improved convolutional neural network is constructed by a plurality of feature channels according to a preset multi-level feature matrix reusing method; according to the shell contour, obtaining shellfish area parameters and a percentage of shellfish damage to the total area of shellfish of the to-be-graded bivalve shellfish; and according to the shellfish area parameters and the percentage of shellfish damage to the total area of shellfish, performing grading processing on the to-be-graded bivalve shellfish. According to the invention, a size and a damaged area of a shell are identified by using machine vision, the damage to shellfish is reduced, damaged shellfish are graded, various limitations of manual measurement are avoided, and the bivalve shellfish grading accuracy is improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a bivalve mollusk grading method and device. Background technique [0002] With the continuous expansion of the production scale of marine shellfish products, it is necessary to classify and test the shellfish in the links of breeding, transportation and sales, because different grades of marine shellfish products have different market values ​​and nutritional values. Big difference. [0003] In the existing bivalve mollusc breeding and production process, the shellfish size is mainly used for multiple classifications. Among them, in the shellfish seedling stage, the sieves are used for manual sieving, and the intermediate seedlings are classified by a classifier. sieve. Yet existing two kinds of methods, all can make shellfish shell be subjected to mechanical vibration and collision in grading work process, vibration can affect the growth and development of shellfish...

AI Technical Summary

Problems solved by technology

Yet existing two kinds of methods, all can make shellfish shell be subjected to mechanical vibration and collision in grading work process, vibration can affect the growth and development of shellfish, and the edge that collision can damage shellfish shell, if shell fragmentation makes mantle Exposed to the outside of the shell will lead to sick and dead shellfish

Moreover, the classification accuracy of the existing classifier is low, the labor intensity of manual screening is high, and the classification efficiency is low

Images