摘要
针对镨\钕(Pr\Nd)萃取过程中元素组分含量难以在线检测的问题,依据Pr、Nd混合溶液图像特征可以反映元素组分含量分布的特性,建立了基于离子颜色特征的Pr\Nd萃取过程元素组分含量预测模型。采集Pr\Nd萃取过程中稀土混合溶液图像信息,在HSI颜色空间中提取图像特征,采用主成分分析法分析各颜色分量对元素组分含量的影响,选取影响较大的H、S特征分量一阶矩作为预测模型的输入变量,利用最小二乘支持向量机算法(LS_SVM)具有解决小样本、非线性能力强及运行速度快的优点,建立基于Pr\Nd萃取过程元素组分含量预测模型。通过Pr\Nd萃取生产现场运行数据测试,表明建立的模型适用于具有离子特征颜色的稀土萃取过程组分含量在线快速预报。
In view of the online detect problem of element component content in Pr/Nd extraction process, according to the property that image features of the mixed solution of Pr and Nd can reflect the distribution character of element component content,a prediction model of component content in Pr/Nd extraction process based on ion color features.At first,collecting image information of rare-earth mixed solution in Pr/Nd extraction process and extracting the image-features in HSI color space.Then,using principal compo-nent analysis method to analyze the impact of each color component on component content and selecting the first moments of H and S features which have more impact on it to the input variables of forecast model.Fi-nally,a prediction model of element component content in Pr/Nd extraction process is established based on least square support vector machine algorithm,it has advantages in solving small sample,non-linear and high speed.An actual data testing shows that the proposed mode can be suitable to online fast forecast of component content in rare-earth extraction process with ion characteristic color.
出处
《南昌大学学报(理科版)》
CAS
北大核心
2013年第6期589-593,共5页
Journal of Nanchang University(Natural Science)
基金
国家自然科学基金项目(51174091
61364013)
江西省自然科学基金资助项目(2010GQS0032)