Research on quantitative fish segmentation method based on line laser scanning

Abstract

Objective： To realize the mechanization of quantitative cutting of fish before processing fish products. Methods: Obtained the fish contour data by linear laser scanning， the distribution of fish body weight with body length was obtained. The quantitative cutting models of fish body with fixed weight and equal weight were constructed, and the validity and accuracy of the method were verified by experiments. Results: The results showed that the predicted curve of fish body weight distribution was close to the real distribution curve, the accuracy of 10 fish was above 91%. The mean absolute errors were 0.16, 0.38 and 1.15 g, respectively. For a given segment weight of 10, 15 and 20 g, the mean absolute errors of 10, 15 and 20 segments were 2.44, 1.35 and 0.67 g, respectively. Conclusion： The research results provide a feasible idea for quantitative segmentation processing of fish, and provide a theoretical reference for the development of intelligent fish processing machinery.

Keywords

• linear laser scanning
• weight distribution
• constant weight segmentation
• equal weight segmentation
• intelligent processing