A New Study Of Starch Phase Transition Using Artificial Intelligence Monitoring System

Being one of the most abundant natural polymers,starch and its derivatives have been widely used in various industrial fields,such as food,medicine,textile and chemical engineer,et al.Phase transition is closely related to components supramolecular structure,and the relative relationship of starch-based system.A deep of phase transition reasonably provides prediction and theoretical guidance for the development of starchy products.Microscopy observation is a simple method and widely used to evaluate gelatinization feature through measuring and recording morphology change of starch granules during heating process.In the past years,though a few techniques with microscopy observation have been positively designed to evaluate the process of starch phase transition,most of these visual observations were marked by subjective uncertainties,and long and costly litigation over discovery.Besides,such methods could not provide online detailed study of granules morphology and phase transitions at various specific points.A combination of quantitative analysis indicators with intelligent analysis methods is important for a deep understanding of starch structure and phase transition.In the present study,an artificial intelligence method including Neural Networks,edge detection and morphological operation,was developed with traditional microscopy observation,and applied to achieve a precise investigation on the changes of granule morphology and starch structure in real time during heating process.Moreover,a DG control system was also designed and used to control the react degree of starchy system.Results showed that:1.An efficient method of automatic object detection combining Canny edge detection with mathematical morphology was developed to quantitatively study the swelling behavior of starch particles with microscopy observation.In the course,swelling capacity(SC)was studied according to the area change of all granules.2.An approach using Artificial Neural Networks(ANNs)method was designed to investigate starch gelatinization based on changes of birefringence during the process of gelatinization.Compared with traditional methods that performed by human operation,which is time-consuming,tedious and subjective,the proposed Starch-SSD is efficient and much faster,and provides a unified standard without subjective uncertainty.3.A gelatinization degree control system,with a combination of ANNs and computer vision,was successfully developed.An intelligent measurement framework was purposely designed to achieve a precise investigation on phase transition and morphology change of starch in real time,as well as a process control during gelatinization.Base on a variation of birefringence number,the degree of gelatinization(DG)control system provided a direct and fast methodology without subjective uncertainty in studying starch gelatinization.