Research On The Food Fineness Perception Based On The Bionic Distributed Mechanical And Intelligent Detecting Simulation System

The artificial intelligence has a rapid development at present,which has emerged in the field of the intelligent food sensory evaluation.At present,the acquisition of sensory information is basically done manually by the assessor.However,relying on assessors for sensory evaluation has certain subjectivity.With the continuous advancement of artificial intelligence technology,other emerging bionic artificial intelligence detection methods such as texture analyzers are more in line with the actual physiological structure of human beings,providing an intelligent alternative approach to sensory evaluation in the field of texture perception.Currently,there are studies about food perception in terms of crispiness,springiness,chewiness and so on,but fineness perception is poorly studied or defined.The distributed mechanical information generated on the flexible tongue is the main contributor to the exploration of fineness perception.However,the detecting indenters of the texture analyzer are mostly simple geometric structures such as a cylindrical shape and a conical shape,which is difficult to reflect the small density felt by the flexible tongue.Therefore,it is of great importance for exploring the fineness perception of food texture to construct in vitro a bionic tongue distributed mechanical testing simulation system.In this thesis,to obtain a texture perception that is closer to the human sense,a bionic tongue distributed mechanical testing simulation system was built by using some advanced technology such as,3D reconstruction of bionic tongue and flexible tactile sensor.Several kinds of food were detected by this system,a fineness perception evaluation model was constructed,proving the feasibility of the system for simulating tongue pressure sensing.Firstly,by capturing a large number of tongue movements during chewing,we analyzed and simulated four representative tongue movement states including the tiled state,sunken state,raised state,and overturned state of the tongue.By analyzing curvature parameters and the Gauss curvature of the tongue surface,we selected the regional circle of interest where the curvature distribution changes the most dramatically and design eight bionic tongue indenters.A flexible arrayed pressure sensor capable of simulating the flexible detection surface for the tongue-palate compression was selected.Together with a software testing platform,we set up a bionic tongue distributed mechanical testing simulation system.Secondly,we selected several kinds of foods as test materials to do the pressure-touch test by using the bionic tongue distributed mechanical testing simulation system.Based on the distributed mechanical information perceived by the surface of the bionic tongue indenter,we established a food fineness perception evaluation model by defining three indicators,including gradient,stress change rate and areal density.The correlation between the sensory assessment and model result was analyzed.The results showed that a significant correlation was achieved.Lastly,the fineness information of food is obtained intelligently based on the bionic tongue distributed mechanical testing simulation system,and displayed based on the multi-pattern recognition,which included support vector machine(SVM),random forest(RF).The optimal accuracy rate is 100%,83.33%,respectively,which shows a good effect of classification for the food texture.It proves the validity of the fineness feature data,and also shows the feasibility of the bionic tongue distributed mechanical testing simulation system for food fineness perception.Then,in order to avoid data redundancy caused by the fineness information,we were committed to the research of exploratory data analysis(EDA),which was used to explore the influence of classification performance for the variable accumulation of the bionic tongue indenter with different curvature distribution.The results show that the method could select different feature combinations aiming at different classification model,which effectively avoids the coverage or hiding of texture information caused by over-learning.In summary,the food fineness evaluation model was constructed based on the bionic tongue distributed mechanical testing simulation system,which was realized the simulation of the fineness perception.