Development of realistic food microstructure considering the structural heterogeneity of cells and intercellular space

The geometric characterization of the microstructure is required to model the micro level transport process inside the plant-based food material during thermal processing. However, the existing geometric construction methods are unable to consider the distinctive geometrical characteristics of cells and intercellular spaces during generation of the virtual microstructure. The main aim of this study is to develop a new method to build microstructure of the plant-based food materials with clear identification of cell and intercellular spaces. Granny smith and red delicious apple were taken as samples to implement and validate the proposed method. The virtual microstructure was generated from the scanning electron microscope image (SEM) of the apples. A new algorithm was developed based on the modified ellipse fitting method to generate the microstructure from microscopic images. Linear Discriminant Analysis (LDA) was used to classify the cells and the intercellular spaces. Statistical analysis was performed between the developed geometrical model and microscopic image to determine the accuracy of the proposed method. The area distribution of the microstructural model found in this study is significantly better than that found in the literature.