Elsevier

Aquacultural Engineering

Volume 63, December 2014, Pages 62-71
Aquacultural Engineering

Review
Visual quality detection of aquatic products using machine vision

https://doi.org/10.1016/j.aquaeng.2014.10.003Get rights and content

Highlights

  • Machine vision is used for visual quality detection of aquatic products.

  • We review visual quality detection based on size and shape.

  • We review visual quality detection based on color.

  • Machine vision is a promising and potential method for automated detection in aquaculture.

Abstract

Aquatic products are popular among consumers and their visual quality used to be detected manually for sorting, grading, species classification and freshness assessment. Machine vision, as a non-destructive method, has been used in external quality detection of aquatic products for its efficiency, objectiveness, consistency and reliability. Quite a number of researches have highlighted its potential for visual quality detection of fishes, fish filets and some other aquatic products (i.e. shrimp, oyster, and scallop). This review introduced detecting methods based on measurement of size, shape, and color using machine vision systems. Size measurement (i.e. length and area) was usually taken for sorting and grading, while shape was measured for species classification with the integration of size information. Color information was studied for analysis of fish filets, fish muscle, fish skin and shrimp, and for color changes of specially treated fish. Machine vision systems used for measuring size, shape, and color was described, including improvements of cameras, illumination settings, image processing and analysis methods, and experimental results as well. With the development in these areas, machine vision technique may achieve higher accuracy and efficiency, and wider application in visual quality detection of aquatic products. Besides, advantages and limitations of these machine vision systems were discussed, with recommendation on future developments.

Introduction

Nowadays, aquatic products have received great popularity because of their high nutritive value and delicious taste. When consumed, their quality would determine their value, price and “best-used-before” date (Sun, 2011). The quality may be presented as appearance, odor, flavor and texture. Appearance attributes, such as size, shape and color, are assigned to visual quality (Alasalvar et al., 2011). These attributes directly influence the products’ acceptance and thus affect most consumers’ purchase trends consciously or subconsciously. Therefore, detecting these attributes is of great significance for better purchase decision and higher economic value.

Traditionally, visual quality detection is predominantly done by trained inspectors, which is labor-costing, time consuming, and difficult to quantify (Balaban et al., 2008). Manual processing and grading is inevitably influenced by human factors such as mistakes, occasional omission in processing as well as fatigue (Mathiassen et al., 2006). Alternatively, machine vision can provide fast, objective and robust measurement (Brosnan and Sun, 2002). Machine vision has been widely applied for quality assurance purposes in different industries (Gümüş et al., 2011). While, the aquaculture industry is a low technology industry for most activities started out in the natural way (Balchen, 1986). Machine vision still has not achieved a common utilization in aquaculture (Zion, 2012). This review described applications of this technology in visual quality detection of some aquatic products, including fishes, shrimp, oyster and scallop. These aquatic products have been studied by many researchers through measuring size, shape, and color parameters using various machine vision systems and methods. The objective of this review is to highlight development in detection methods, machine vision systems, image analysis and processing approaches, and analyzes its characteristics, so as to unlock the potential application of machine vision in aquaculture.

Section snippets

Machine vision system

Machine vision is a novel technology for recognizing objects, extracting and analyzing quantitative information from digital images. A typical machine vision system (MVS) often consists of an image acquisition system, image processing and statistical analysis procedures as shown in Fig. 1. Essential elements of an image acquisition system include: a camera, an illumination device, a frame-grabber, and a computer. Images are processed via pre-treatment, segmentation, and feature extraction (

Visual quality detection based on size and shape measurement

Visual quality detection, for the seek of sorting, grading, counting and species classification of aquatic products, was mainly based on measurement of size (i.e. length, area, volume, width, height and so on) and shape. In this section, methods and systems for size and shape measurement were reviewed for several application fields including fishes, shrimp, scallop and oyster.

Visual quality detection based on color analysis

Color is an important visual quality criteria of many products. It contains the basic information corresponding to human vision (Dowlati et al., 2012) and is closely related to consumer perception. Consumers usually tend to associate color with freshness, high quality and better flavor (Gormley, 1992). Color can be measured subjectively by sensory panels, which, however, is difficult to convert to reproducible numerical values (Balaban, 2008). Color can also be determined by instrumental

Discussion and conclusions

This paper reviewed the applications of machine vision and associated technologies in visual quality detection of aquatic products. After years of development, machine vision technology has been improved along with progress of cameras, illumination systems, design of the light box, and also image analysis methods. Parameters used for size measurement ranged from length to area, perimeter, width and heights on the purpose of sorting and grading, and species classification. For color measurement,

Acknowledgements

We thank the National Key Technologies R&D Program of China (2012BAD29B04) for funding of this project. The research was supported by the Program for Zhejiang Leading Team of S&T Innovation (2011R50029).

References (79)

  • A. Kohler et al.

    Sorting salted cod filets by computer vision: a pilot study

    Comp. Electron. Agric.

    (2002)
  • C.A. Martı́nez-Palacios et al.

    Effect of temperature on growth and survival of Chirostoma estor estor, Jordan 1879, monitored using a simple video technique for remote measurement of length and mass of larval and juvenile fishes

    Aquaculture

    (2002)
  • P. Marty-Mahé et al.

    Quality traits of brown trouts (Salmo trutta) cutlets described by automated color image analysis

    Aquaculture

    (2004)
  • A. Mateo et al.

    Quality analysis of tuna meat using an automated color inspection system

    Aquacult. Eng.

    (2006)
  • B.P. Ruff et al.

    Fish sizing and monitoring using a stereo image analysis system applied to fish farming

    Aquacult. Eng.

    (1995)
  • A.C.R. Shieh et al.

    Measurement of fish size in Atlantic salmon (Salmo salar L.) cages using stereographic video techniques

    Aquacult. Eng.

    (1998)
  • F. Storbeck et al.

    Fish species recognition using computer vision and a neural network

    Fish. Res.

    (2001)
  • N. Strachan

    Length measurement of fish by computer vision

    Comput. Electron. Agric.

    (1993)
  • R. Tillett et al.

    Estimating dimensions of free-swimming fish using 3D point distribution models

    Comput. Vis. Image Und.

    (2000)
  • G. Xiong et al.

    Shape similarity measure using turn angle cross-correlation for oyster quality evaluation

    J. Food Eng.

    (2010)
  • D. Zhang et al.

    Automatic shrimp shape grading using evolution constructed features

    Comput. Electron. Agric.

    (2014)
  • B. Zion

    The use of computer vision technologies in aquaculture – a review

    Comput. Electron. Agric.

    (2012)
  • B. Zion et al.

    Real-time underwater sorting of edible fish species

    Comput. Electron. Agric.

    (2007)
  • B. Zion et al.

    Sorting fish by computer vision

    Comput. Electron. Agric.

    (1999)
  • B. Zion et al.

    In-vivo fish sorting by computer vision

    Aquacult. Eng.

    (2000)
  • H. Aguirre et al.

    Length–weight relationship for 21 fish species from a coastal lagoon in the southwestern Gulf of California

    J. Appl. Ichthyol.

    (2008)
  • C. Alasalvar et al.

    Handbook of Seafood Quality, Safety and Health Applications

    (2011)
  • M.K. Alsmadi et al.

    Fish recognition based on robust features extraction from size and shape measurements using neural network

    J. Comput. Sci.

    (2010)
  • H. Arnarson

    Fish and fish product sorting

    Fish Quality Control by Computer Vision

    (1991)
  • M.Ö. Balaban

    Quantifying nonhomogeneous colors in agricultural materials. Part I: Method development

    J. Food Sci.

    (2008)
  • M.Ö. Balaban et al.

    Quantifying nonhomogeneous colors in agricultural materials. Part II: Comparison of machine vision and sensory panel evaluations

    J. Food Sci.

    (2008)
  • M.Ö. Balaban et al.

    Color enhancement and potential fraud in using CO

    Modified Atmospheric Processing and Packaging of Fish: Filtered Smokes, Carbon Monoxide, and Reduced Oxygen Packing

    (2007)
  • M.Ö. Balaban et al.

    Using image analysis to predict the weight of Alaskan salmon of different species

    J. Food Sci.

    (2010)
  • M.Ö. Balaban et al.

    Determination of count and uniformity ratio of shrimp by machine vision

    J. Aquat. Food Prod. Technol.

    (1995)
  • J.G. Balchen

    Bridging the gap between aquaculture and the information sciences

    Model. Identif. Control

    (1986)
  • C. Costa et al.

    An advanced colour calibration method for fish freshness assessment: a comparison between standard and passive refrigeration modalities

    Food Bioprocess Technol.

    (2012)
  • C.I. de Ciencias Marinas et al.

    Length-weight relationship of demersal fish from the Eastern coast of the mouth of the Gulf of California

    J. Fish. Aquat. Sci.

    (2010)
  • K. Diehl et al.

    Geometric and physical properties of raw oyster meat as related to grading

    Trans. ASAE

    (1990)
  • J.P. Frisby et al.

    3D Model Recognition from Stereoscopic Cues

    (1991)
  • Cited by (0)

    View full text