摘要
脐橙的检测方法目前主要还是采用劳动强度大、工作效率低、随意性大、客观性不强、也不符合当前标准化要求的人工检测方法。运用计算机视觉和模式识别技术,研究了基于计算机视觉的脐橙分级系统,首先对获取的图像分割出背景、脐橙本体和表面缺陷,再根据脐橙的分级标准,提取出果实横径、表面缺陷的特征参数,采用径向基神经网络对脐橙样本进行了等级识别,实现脐橙的自动检测与分级。
The navel orange＇s detection method at present has many defects, for example, it results in physical or mental fatigue of human inspectors, so it can not guarantee the objectivity of detection, and the detection can not always be implemented with great efficiency and high speed, and it can not meet the requirement of social legal system and standardization. According to the characteristics and requests of navel orange, this study uses computer vision and pattern recognition technology, segmenting the background, navel orange body and its surface defects, then the parameters of the diameter and the surface defects are collected, and realizes the automatic detection and classification.
出处
《江西农业大学学报》
CAS
CSCD
北大核心
2006年第2期304-307,共4页
Acta Agriculturae Universitatis Jiangxiensis
基金
国家自然科学基金资助项目(30460059)
江西农业大学青年科学基金项目(1332)
关键词
脐橙
计算机视觉
图像处理
模式识别
神经网络
navel orange
computer vision
image processing
pattern recognition
neural network
