Feasibility study of detecting plum’s early mechanical injury based on optical coherence tomography and cell morphological parameters

Abstract

Mechanical injury, which is hardly observed by the naked eye and machine vision technique if the epidermis of the plum is not ruptured, tends to be the main cause of the shortened shelf life and decay of plums. For the purpose of non-destructive discrimination of the mechanical injury, this study proposes to apply optical coherence tomography (OCT) to conduct 3D observation of the plum's subcutaneous tissues and to automatically compute the tissues' morphological parameters. The OCT image processing consists of image denoising, binarization, cell segmentation, region selection, and calculation of cell morphological parameters. Afterwards, mechanical injury can be detected based on the morphological parameters. During the experiment, seven 3D parameters including equivalent diameter, the longest Feret diameter, the shortest Feret diameter, anisotropy, the total cell surface area, the total cell volume, and the total cell amount are figured out from OCT data. Mechanical injury results in damage of cell integrity, a sudden increase of cell equivalent diameter, longest Feret diameter, and shortest Feret diameter, which are decreased after the peak value, as well as an obvious reduction of cell total surface area, volume, and amount. All the changing regularities of morphological parameters lay the foundation for detecting the plum's early mechanical injury. Novelty impact statement OCT acquires 3D of sub-surface structure of plum and is tried as a non-destructive method for plum injury. Parenchyma cells are selected from 3D images, and their morphological parameters are figured out by the proposed method. Some cell morphological parameters can be used for mechanical injury discrimination.