Research On Key Technologies Of Fish Freshness Detection Based On Impedance Spectroscopy

Traditional physical and chemical testing for meat freshness hastedious operation,long inspection cycle,and certain requirements for the professional level of the operators and the hardware conditions of the laboratory.Electrochemical impedance spectroscopy technology(EIS)has many advantages,like high speed,non-destructive,low cost and portability.It has broad application prospects in the fields of agricultural engineering,food science,biotechnology.At present,EIS for fish freshness evaluation can achieve good results in the laboratory research with the same sample origin and stable detection environment.However,the same accuracy of detection is difficult to to be achieved by the same measurement method under the practical application scenario.Three key problems restrict the further development of EIS.1)The effect of individual differences on the test results.2)The effect of ambient temperature on the measurement results.3)Most of the existing freshness assessment models are based on regression analysis.Due to the lack of theoretical support that can reflect the inherent mechanism and essential law,the model is poor in universality,and difficult to effectively guide the experiment or data analysis.Aiming at these three key problems,this paper explores from the aspect of establishment of experimental platform,optimization of parameters in measurement system,extraction and selection of characteristic parameters,construction and evaluation of classification model,modeling for the temperature compensation,temperature correction of classification model.An effective solution is integrated to reduce the impact of individual differences and environmental temperature on EIS measurement results.The main innovative works of this paper are summarized as follows:1.According to the influence which is resulted from different electrical characteristics of fish from different sources on impedance spectroscopy,parameters of orthogonal anisotropy and morphological characteristic were proposed respectively as the index of freshness evaluation.Compared with the traditional classification model,the classification accuracy of the model based on orthogonal anisotropy characteristic parameters is improved by 22-25 percentage points.Based on the basic theory of bioimpedance spectroscopy,the mechanism for parameters of characteristic was discussed,and the parameters of morphological characteristic were found that the physical meaning of constant phase angle coefficient obtained from the morphological characteristic parameters is similar to the papers in-depth mining on the basis of the conventional equivalent elements of circuit.Essentially,it is related to the relative changes of cell components with resistance and characteristics capacitance in the process on spoilage of fish.2.Aiming at the problems of sensitivity to quantity of sample,lack of theoretical support and poor universality of modeling methods based on regression analysis and data fitting,a modeling method based on the intrinsic correlation between parameter of characteristic on impedance spectrum curve and fish spoilage stage was proposed.Compared with the former ones,the classificationof freshness model constructed by this method improves the classified accuracy of samples from complex sources by 28-31 percentage points.The modeling method of non-data fitting establishes the relationship between the changes in macro-electrical properties of samples and the changes in microstructure of tissues and cells in the course of corruption,and makes a preliminary attempt to promote the study of bioimpedance spectroscopy from level of the statistical law to the level of internal mechanism.3.A multi-parameter impedance spectroscopy fusion of classification on fish freshness model based on model similarity is proposed.In order to eliminate the influence of subjectivity and uncertainty on the fusion results,the differences among the parameter of characteristic models of impedance spectra are quantified and used as the basis for determining the weight coefficients of fusion parameters.Compared with the single evaluation index,the accuracy of classification on the degree of freshness from complex source samples was improved by 9-12 percentage points by multi-parameter fusion model.4.In order to reduce the impact of ambient temperature on results of classification on the degree of freshness,a model for compensation of temperature was constructed.The model is fused with the algorithm for optimal classification on the degree of freshness selected in the previous paper.Compared with the classification results without compensation of temperature,the accuracy of classification on the degree of freshness was improved by 38 percentage points after temperature correction.The results show that in the field of online monitoring,the model of classification on the degree of freshness based on the mutation point at the curve of morphological characteristic is the best.In the field of real-time detection,model of classification based on feature at parameters of difference among orthogonal direction is the best.Temperature correction can significantly improve the accuracy of classification on the degree of freshnessin field detection with variable ambient temperature.The research ideas and solutions to the key problems affecting the freshness detection of fish by impedance spectroscopy are put forward in this paper.It also has certain reference value for the research on other technology of object-oriented impedance spectroscopy biological detection.