RBF神经网络在土壤重金属污染评价中的应用

Pollution Assessment of Heavy Metals in Soils Based on RBF Neural Network

运用MATLAB软件的K均值聚类算法结合神经网络工具箱,通过建立RBF神经网络对郑汴路圃田—杏花营段路旁土壤进行了评价。结果表明,土壤重金属积累峰值污染峰值多出现在距路基50m范围以内,并随着距路基距离的增加,污染程度逐渐下降。风力因素和邻近铁路对路段两侧土壤重金属污染影响较大,且公路运营时间越早,对两侧土壤污染越严重。杏花营断面土壤质量以Ⅱ级为主,北侧断面土壤质量多为Ⅲ级。圃田断面各样点综合评价结果均为Ⅱ级,整体质量优于杏花营断面。RBF神经网络评价方法具有计算速度快,评价结果客观等优点。

RBF network which was established by using K-means algorithm of MATLAB and combing with neural network toolbox was used to assess soil quality of Putian-Xinghuaying Section,Zhengzhou-Kaifeng Highway. Results showed that peak of heavy metal contamination appeared in the range of 50m away from the roadbed,with Cd and Zn content in the high level. With distance increasing,the degree of pollution decreased gradually. Degree of heavy metal pollution on both sides of the Highway was affected by wind and adjacent railway seriously. The transport of road operator earlier,soil pollution will be more serious. Soil quality of Xinghuaying Section was mainly in Grade II,and most parts of north side were in Grade III. Comprehensive evaluation of various points about Putian Section was in Grade II. The method is advantageous over faster calculation and objectively assessment.