Detection Of Fish Quality Based On Hyperspectral Imaging Technology

Fish plays an important role in people's daily life as if it can provide abundant proteins,unsaturated fattyacid,vitamins and minerals.Fish is succulent with high moisture content,easily in fish due to the activity of catabolic enzymes and microorganisms.Therefore,it is significant to detect fish quality.However,traditional methods for quality are destructive,tedious,time-consuming,thus they can no longer meet the requirements of rapid,non-destructive and real-time measurement for modern digital fishery development.This study applied spectroscopy and hyperspectral imaging technique,combined with image processing methods and chemometric methods to inspect the quality and safety of marine fish.Three parts are included: the discrimination of frozen-thawed and fresh turbot,the mapping of fat contents in salmon,development of the instrument for detecting fish freshness by hyperspectral imaging technique.The main research contents and results are shown as follows:(1)The fresh and frozen-thawing fish was identified by the visible near infrared hyperspectral imaging technique.The freezing temperature is 20 C below zero,and the number of thawing is 1 times,freezing thawing and thawing,respectively,2 times.Extraction of fish samples of hyperspectral images of interest area spectrum,using genetic algorithm(GA),competitive adaptive reweighted algorithm(CARS),genetic algorithm(SPA)feature extraction of wavelength,the wavelength of these variables is set up based on soft independent modeling of class analogy(SIMCA)and least squares support vector machine(LS-SVM)identification model.The best overall correct discrimination rate was98%.(2)Long wave near infrared hyperspectral imaging system for the salmon fat content detection and visualization based on distribution.The original spectral data is corrected,and the region of interest(ROI)is extracted to extract the sample spectral data.The continuous projection algorithm(SPA)is used to extract the characteristic wavelengths,and input the linear PLSR model.The coefficient of the fat prediction set is 0.913,and the root mean square error of prediction is 0.921.Finally,based on pseudo color expression to predict salmon fat content and graphical visualization.(3)The detection platform for detecting the freshness of fish by hyperspectral technology was studied and used to identify the freshness of turbot.It is made up of screw,motor,dark box,image acquisition module,control device,halogen lamp,fish sample,bracket,control unit,adjustment seat,etc.