	US 12,236,693 B2
	Method for non-destructive ripeness identification of kiwifruit based on machine vision learning
Shengbing Xu, Guangdong (CN); Xiaoquan Cai, Guangdong (CN); Zhenyou Wang, Guangdong (CN); and Jinzhang Li, Guangdong (CN)
Filed by GUANGDONG UNIVERSITY OF TECHNOLOGY, Guangdong (CN)
Filed on Aug. 12, 2022, as Appl. No. 17/886,701.
Claims priority of application No. 202111510664.9 (CN), filed on Dec. 11, 2021.
Prior Publication US 2023/0186656 A1, Jun. 15, 2023
Int. Cl. G06V 20/68 (2022.01); G06V 10/54 (2022.01); G06V 10/56 (2022.01); G06V 10/77 (2022.01)

CPC G06V 20/68 (2022.01) [G06V 10/54 (2022.01); G06V 10/56 (2022.01); G06V 10/7715 (2022.01)]

8 Claims

1. A method for non-destructive ripeness identification of kiwifruit based on machine vision learning, comprising:

S1: collecting kiwifruit data to obtain an original data set by collecting images of 40-80 kiwifruits in a same period of time over 3-6 days, recording a label, which comprises ripeness information obtained by pressing at a the same location of a corresponding kiwifruit to determine whether the corresponding kiwifruit is ripe using an empirical judgment method, for each of the images, and saving each of the images with the label;

S2: extracting a the color and a the texture of a kiwifruit skin from each of the images in the original data set; and

S3: training a deep learning model to learn a connection between the color and the texture of the kiwifruit skin and the ripeness information of the corresponding kiwifruit, wherein the ripeness information is divided into three stages-unripe, slightly ripe and ripe, using the color and the texture of the kiwifruit skin extracted from each of the images and the label.