面向环境监测的传感器网络边缘检测研究

Research on edge detection of sensor network for environmental monitoring

为了更好地了解地区空气中污染物的气体浓度和分布情况,更有针对性地实施环境保护政策,文中提出一种基于风向和风速的空气污染物插值方法,利用该方法进行传感器网络的边缘检测。在此利用唐山市开平区检测站点检测到的2月份数据进行实验,按照相关临近图将检测点组成网络,大致判断出边界节点;再利用基于风速风向扩散模型插值方法对所在区域进行插值,从而较为准确地预测出边界。将所提方法与反距离权重法和基于场强的插值方法相比,在PM2.5浓度较小时,反距离权重法更接近真实值;当PM2.5数值较大时,风向风速插值法更接近真实值,但风向风速插值法的根均方误差为10.35851403,明显低于其他两种方法。由此可见,基于风速风向扩散模型插值方法有效地解决了监测地区监测站点少、流动性差、分布不均的问题,可以准确地预测出污染区域边界,提高检测效率与精度。

In order to better understand the gas concentration and distribution of air pollutants in a region and implement environmental protection policies,an air pollutants interpolation method based on wind direction and wind speed is proposed to conduct the edge detection of sensor networks.The February data detected by Kaiping detection station in Tangshan City is used for the experiment.The network of detection points is formed according to the relevant proximity graph to roughly judge the boundary nodes.The interpolation method based on wind speed and wind direction diffusion model is used to perform interpolation for the region to accurately predict the boundary.In comparison with the inverse distance weight method and the interpolation method based on field strength,the inverse distance weight method is closer to the real value when the PM2.5 concentration is small,and the wind direction and wind speed interpolation method is closer to the real value when the PM2.5 value is larger.The root mean square error of wind direction and wind speed interpolation method is 10.35851403,which is significantly lower than that of other two methods.It can be seen that the interpolation method based on wind speed and direction diffusion model can effectively solve the problems of no enough monitoring stations,poor mobility and uneven distribution in the monitored area,accurately predict the boundary of the pollution area,and improve the detection efficiency and accuracy.