摘要
针对芒果内部腐烂程度难以预测的问题,提出一种在MATLAB中利用机器学习建立芒果缺陷度腐烂度预测模型的方法。首先,采集芒果缺陷图像,并进行滤波去噪平滑噪声。然后,利用迭代阈值分割和形态学运算,提取芒果果皮、果肉、表面缺陷和果肉腐烂图像。最后,提取芒果表面特征,并定义果皮缺陷度及果肉腐烂度,运用BP神经网络进行数据拟合,建立缺陷度腐烂度预测模型。实验结果表明本文建立的预测模型对芒果腐烂度的预测平均准确率达到88.3%。
In view of the problem that it is difficult to predict inner rot degree of mango,a method is proposed to establish a prediction model of mango defect degree-rot degree by using machine learning in the MATLAB.Firstly,the image of mango defect is collected,with the smoothing noise removed by filtering.Then,the images of mango peel,flesh,surface defect and flesh rot are extracted by using iterative threshold segmentation and morphological operation.Finally,the surface features of mango are extracted,with the defect degree and pulp rot degree of mango defined.The BP neural network is used for data fitting to establish the defect degree and rot degree prediction model.The experimental results showed that the average accuracy of the prediction model for mango rot reached 88.3%.
作者
张铮
周嘉政
柯子鹏
钱勤建
胡新宇
ZHANG Zheng;ZHOU Jiazheng;KE Zipeng;QIAN Qinjian;HU Xinyu(School of Mechanical Engineering,Hubei University of Technology,Wuhan 430068,China)
出处
《西安理工大学学报》
CAS
北大核心
2022年第2期151-157,共7页
Journal of Xi'an University of Technology
基金
国家自然科学基金资助项目(61976083)。
关键词
芒果
机器学习
缺陷度腐烂度预测模型
机器视觉
BP神经网络
mango
machine learning
defect degree-rot degree prediction model
machine vision
BP neural network
