摘要
为了分析海水中所含有害藻的种类和数量,做到赤潮藻类生物发展的早期监测、预报,开发了一赤潮藻类图像计算机自动识别系统。运用图像处理技术提取藻类图像形态、纹理特征等,运用遗传算法进行特征选择。在此基础上用神经网络建立分类器,对藻类图像进行分类识别。结果表明,该系统能有效提高学习能力和分类性能,对引发赤潮的3种主要藻类达到了很好的分类识别,分析结果与人工计数识别结果相差较少。
According to analyze the species and quantities of the harmful algae in water, and the monitoring and prediction of the red tide ahead, an automatic recognition system for the harmful algae images is made. The image processing technology extracts the harm- ful algae images features, texture features, etc. The method is proposed to feature selection based on the genetic algorithm. Finally, it could be used the neural network classifier. The harmful algae images can be classified and recognized. The results showed that the system can improve the ability of study and recognition. The classified results have a little difference compared with the artificial method.
出处
《海洋环境科学》
CAS
CSCD
北大核心
2007年第1期42-44,共3页
Marine Environmental Science
基金
国家自然科学基金(50447008)
关键词
赤潮藻
特征选择
遗传算法
神经网络
harmful algae
feature selection
genetic algorithm
neural network