摘要
为了有效提高水果分级系统的分类正确率,利用图像处理、模式识别,以及神经网络分类器融合等技术,构造了一个水果等级分类系统。以3个最具代表性的水果的外在品质特征作为神经网络的输入,并融合各神经网络分类器的结果对水果进行自动化分级。实验结果表明:系统分类效果较好,可用于水果深加工生产。
To improve accuracy of the of fruit quality classification, a system is constructed using image processing, pattern recognition and neural networks. Three most representative fruit surface quality conditions obtained from images are used as input to neural networks. Classification results of the neural network classifier are fused to separate the fruits into different quality levels. Experiments show that the system can produce good results of fruit classification. The method can be used in the deep processing of fruits.
出处
《上海电机学院学报》
2012年第3期163-166,170,共5页
Journal of Shanghai Dianji University
基金
上海市大学生科技创新项目资助(2011SCX36)
关键词
外在品质
特征提取
神经网络分类器
多分类器融合
surface quality condition characteristic extraction neural network classifier fusion of multi-classifier
