Application Of Pattern Recognition In Food Authentication

As people's expectations for quality of life are getting higher and consumers are also demanding higher quality foods.Therefore,food authentication technology has also received the attention of the scientific community.Food authentication is a technical approach to ensure the nature of food.The main authentication method is based on spectroscopy,which has the characteristics of fast information collection and non-destructive testing.However,the limitation of this technology is that it is expensive and at the same time It requires a professional experimental environment and personnel operations,and it is difficult to promote it to the market.Therefore,it is very important to research and develop a new,low-cost,fast and efficient food authentication system.At present,computer vision is also one of the most popular food authentication technologies.It extracts the physical features of food by pattern recognition method.The technology has the characteristics of non-contact,high speed and high precision.Traditional computer vision technology requires careful setup of external cameras and lighting systems to ensure that the required accuracy is up to standard.These limitations make computer vision technology unsuitable for consumers.In addition,traditional computer vision technology does not provide spectral information that is critical to the composition of the food.In this paper,we take apple,milk and olive oil as research objects,and uses the diffraction grating,computer vision technology and pattern recognition algorithm to preprocess,model and analyze the data.Modeling algorithms such as partial least squares discriminant analysis(PLS-DA),kernel partial least squares(PLS),and local weighted partial least squares(LW-PLS)are used to model the diffraction grating and the proposed machine.The feasibility of visual methods in food certification,and provides new ideas for food authentication research.1.Food authentication based on diffraction grating(1)Design a computer vision system based on diffraction grating,take the organic apple authentication as the research object,carry out data acquisition,preprocessing,and establish an organic apple authentication model based on pattern recognition algorithm.The accuracy rate of organic apple authentication is 93.3%.A new research solution for future research on new low-cost,fast and non-destructive testing solutions.(2)In the organic apple authentication system based on diffraction grating,establish 12 pattern recognition and discriminant models based on partial least squares(PLS),support vector machine(SVM)and random forest(RF),and explore suitable pattern recognition algorithm for food authentication.2.Food authentication based on computer vision technology(1)Establish an organic apple authentication model by using computer vision technology combined with pattern recognition algorithm.Using image processing technology,the local binary model(LBP)is used to extract the texture features in the apple image.Through experiments,the accuracy of organic apple authentication can reach 94%.The experimental results show that the method can obtain the experimental standard of professional instruments under the condition of visible light.(2)Using computer vision technology combined with smart phone module to achieve the fat content of milk,rapid determination of olive oil adulteration,develop a new portable liquid food certification system for data collection of experimental samples,the system uses the spectral spatial information of the image Modeling,while extending the local partial least squares(LW-PLS),using this algorithm to quickly achieve data classification,olive oil adulteration accuracy can reach 96.2%,milk fat content measurement accuracy is up to 100 %.It illustrates the feasibility of the system to be promoted to the mass market.Research shows that the proposed food authentication method based on diffraction grating and computer vision is feasible and can be promoted to the mass market,and It provides an effective solution for future food authentication research.