摘要
阐述了在储粮害虫分类识别研究中利用计算机数字图像处理技术,自动提取静态储粮害虫图像的数理统计特征、纹理特征和几何形状特征,并在此基础上采用BP神经网络进行分类和识别的主要技术和方法.实验结果表明,BP具有较强的自适应性,对有噪声、残缺的储粮害虫图像识别也能得到较好的效果.
<Abstrcat> In the classification of grain pest, the Computer-aided digital image processing technique was used to extract the mathematical and statistic eigenvalue, the texture eigenvalue and geometrical configuration features of the pests. The technique and method of using BP neural network in the classification of grain pest was researched. The result indicated that BP has the stronger self-adaptable ability and it can revise the noisy and the damaged grain pest images very well.
出处
《河南工业大学学报(自然科学版)》
CAS
北大核心
2005年第1期19-22,共4页
Journal of Henan University of Technology:Natural Science Edition
基金
河南省科技攻关计划项目(0224010011)
关键词
储粮害虫
图像增强
特征提取
模式识别
神经网络
grain pest
image enhancement
feature extraction
pattern recognition
neural network