@Article{JFBI-7-441,
author = {Aixia Guo, Deqin Xiao and Xiangjun Zou},
title = {Computation Model on Image Segmentation Threshold of Litchi Cluster Based on Exploratory Analysis},
journal = {Journal of Fiber Bioengineering and Informatics},
year = {2014},
volume = {7},
number = {3},
pages = {441--452},
abstract = {To construct a litchi harvesting robot, the first key part is the machine vision system which is used to
recognize ripe litchi clusters and their main fruit bearing branch. It selects and locates picking points.
Hence, in order to establish a threshold computation model used to recognize litchi cluster, the research
focus is in recognizing all parts of the litchi image. In this paper, a procedure on how to develop
an automatic recognition of litchi cluster, fruits and their main fruit bearing branch guided for litchi
harvesting robot is proposed. Firstly, according to the analysis on the specialty of litchi fruits and their
main fruit bearing branch, particularity and uncertainty of illumination and environment, an overall
scheme on the threshold computation model is used to recognize the litchi cluster based on exploratory
analysis and its' application are provided. Secondly, after analyzing and comparing all thresholds,
running time and effect on image segmentation by threshold computation methods of the maximum
entropy, iterative, Otsu and histogram bimodal method, the interval for recognizing all sorts of litchi
clusters is obtained, and a mathematical model for computing threshold to segment litchi cluster is put
forward. Finally, all ripe litchi clusters of testing images from 6 groups (all together 120) of differently
illuminated (in high light, normal light and backlighting) litchi images from differently-colored main
fruit bearing branches (partial red, partial brown and partial brown) collected in natural circumstance
are effectively recognized with the threshold segmentation method based on the given computing model,
with recognition ratio of 88.89%, 92.0%, 88.24%, and 95.45%, 90.0%, 83.33%, which can satisfy the
request of image segmentation on litchi-picking robots in complicated environment.},
issn = {2617-8699},
doi = {https://doi.org/10.3993/jfbi09201413},
url = {http://global-sci.org/intro/article_detail/jfbi/4799.html}
}