Research On Characterization Method And Device For The Flavor Of Fermented Bean Curd Based On Olfactory And Taste Visualization Technology

Fermented bean curd,a traditional fermented soybean product in China,is an important seasoning and accompaniment food in People’s Daily Life,with a distinct flavor and high nutritional value.As one of the important indexes affecting the quality of fermented bean curd,it is of great significance to realize its scientific characterization.At the moment,the characterization methods of fermented bean curd flavor primarily include traditional physiochemical detection,instrument detection,and artificial sensory evaluation.Physiochemical and instrument testing,for example,have the disadvantages of being time-consuming and complicated to perform.The evaluator in artificial sensory evaluation is easily fatigued and it is inconvenient to work for an extended period of time.In order to make up for the shortcomings of traditional detection methods and realize the scientific and rapid characterization of fermented bean curd flavor,The olfactory visualization and taste visualization techniques were used in this study to detect the volatile and taste substances of fermented bean curd,and a visual characterization method of fermented bean curd flavor was established.On this basis,a visual detection device for fermented bean curd flavor was developed.The main research contents and results of this research are as follows:1.Physicochemical detection of flavor substances in fermented bean curdThe volatile substances in fermented bean curd were determined using gas chromatography-mass spectrometry(GC-MS).Using partial least squares discriminant analysis,the detected substances were analyzed,and 13 volatile substances were chosen as the main aroma substances of fermented bean curd.The chemical method SB/T 10170-2007 was used to determine total acid,reducing sugar,salt,and amino acid nitrogen.The results revealed that the content of physicochemical substances differed between fermented bean curds.Based on these findings,the content of physicochemical indexes was used as a reference for developing a quantitative prediction model of characteristic taste substances in fermented bean curd.2.Artificial sensory evaluation of flavor quality of fermented bean curdThe detailed rules for the artificial sensory evaluation of fermented bean curd were developed in accordance with the sensory requirements in the industry standard SB/T 10170-2007.To assess the flavor intensity and flavor quality of various fermented bean curds,the quantitative description method was used.The results revealed that the six types of fermented bean curd had a strong aroma of wine and soy sauce,as well as an obvious aroma of fermentation and ester.The primary flavors are saltiness and umami.Furthermore,the six types of fermented bean curd can be graded using an artificial sensory evaluation method.3.Research on smell detection method of fermented bean curd based on olfactory visualization technologyA sensor array was built to detect the smell information of fermented bean curd using the reaction mechanism of olfactory visualization technology.The K-Nearest Neighbor(KNN)and Support Vector Machine(SVM)models were used to develop the fermented bean curd brand identification model.The two models’discriminant accuracy was 84.72%and 94.44%,respectively.Simultaneously,the KNN model and the SVM model were developed to differentiate the aroma quality grades of fermented bean curd,and the recognition accuracy of the prediction sets of the two models was94.44%and 97.22%,respectively.Furthermore,the Partial Least Square(PLS)and SVM algorithms were used to build quantitative prediction models of fermented bean curd’s main aroma substances.The R_Pof the SVM model prediction set is 0.7437,and the RMSEP is 8.5213,which is better than the PLS model.The results show that olfactory visualization technology can effectively distinguish between fermented bean curd brands and aroma quality grades,as well as quantitatively predict the content of main aroma substances in fermented bean curd.4.Research on taste detection method of fermented bean curd based on taste visualization technologyA taste visualization sensor array was built to collect information about the taste of fermented bean curd.The brand discrimination model of fermented bean curd was established using KNN and SVM,and the discrimination accuracy was 95.83 percent and 97.22%,respectively.To distinguish the taste quality grades of fermented bean curd,KNN and SVM models based on taste visualization technology were also developed.The results show that both models can accurately distinguish the level of taste quality.Based on taste visualization technology,PLS and SVM were used to build a quantitative prediction model of the characteristic taste substance of fermented bean curd.The RP of total acid,reducing sugar,salt,and amino acid nitrogen in the PLS model were 0.9250,0.9035,0.7301,and 0.8591,respectively,and the RMSEP were 0.0848,0.8130,0.2920,and 0.0355.The R_Pof the SVM model was 0.9035,0.9170,0.7716,and 0.9123,while the RMSEP was 0.8934,4.3105,7.5091,and0.5381.The results demonstrated that using taste visualization technology,it was possible to quantitatively predict the characteristic taste substances of fermented bean curd.Except for salt,the prediction results of other indexes were all good.The SVM model established was generally superior to the PLS model.5.Research on visual characterization method and device of fermented bean curd flavor based on olfactory and gustatory fusionhe visual information of smell and taste were combined to establish the KNN model and SVM model for fermented bean curd brand identification.The accuracy rates of the two models are 97.22%and 100%,respectively.The fusion technique produced more accurate results than the single technique.The comprehensive flavor quality grade of fermented bean curd was discriminated using fusion technology,and two grade discrimination models of KNN and SVM were established.Both models achieved 100 percent accuracy.The findings demonstrated that the fusion technology could be used to effectively differentiate the comprehensive flavor grade of fermented bean curd.The fusion technology was also used to develop a quantitative prediction model for fermented bean curd physicochemical indexes.In terms of quantitative prediction of the four taste indexes,the SVM model performed significantly better than the PLS model.In general,the prediction effect of the model based on visual fusion information outperforms that of the model based on single taste information.The hardware and software of the visual detection device of fermented bean curd flavor were designed based on research on flavor characterization methods.The design of the reaction chamber and the selection of image acquisition components were included in the hardware component.The software component includes sensor array image processing as well as the creation of an image processing interface based on Python and OpenCV.The flavor characterization of fermented bean curd is the focus of this research.The odor substances of fermented bean curd were detected using olfactory visualization technology,and the characteristic taste indexes of fermented bean curd were detected using taste visualization technology.The results demonstrated that both techniques were capable of characterizing the flavor of fermented bean curd,and that the visualization technology could not only qualitatively distinguish between different brands of fermented bean curd,but also quantitatively predict the contents of corresponding physicochemical substances.Fusion technology has the potential to improve detection accuracy.The established method can serve as a reference for flavor control in the production of fermented bean curd and flavor control in the finished product,as well as theoretical support for the application of visualization technology in the field of traditional fermented food flavor detection.