Research And Implementation Of Data Correction Method For Meteorological Sensor Network

At present, the meteorological data monitoring mainly depends on the weather stations across the country. Because of the shortcomings of the weather stations, such as low distribution density, single deployment location, poor mobility, high cost, the current meteorological data monitoring model is difficult to meet the requirements of many special applications.Using wireless sensor network technology in meteorology, formatting meteorological sensor network and collecting data by meteorological sensor network, can take the advantages of wireless sensor network, including high density and low cost. In order to achieve rapid deployment, high-density deployment, meteorological sensors in sensor networks tend to choose non-professional sensor, the sensor package while also using inexpensive non-professional package, which brings issues, includes the accuracy of data, availability and other. In this paper, meteorological data correction methods was proposed based on wavelet de-noising, BP neural network. The main contribution of this paper is as follows:(1) A data correction method is presented based on BP neural network. We preprocess the data by the wavelet threshold de-noising algorithm to find the pattern of data. Then, the filtered data as input while standard data as output, we use BP neural network to establish data correction model. Finally, we use this model to correct other observed data. The experiment shows that this method can greatly improve the accuracy of collected data.(2)An improved data correction method is proposed by using the relationship between temperature and solar radiation. We analyze the relationship between hourly solar radiation and the absolute error first. Then, we use the observation data of different radiation, radiation data and standard data to establish the data correction model. The experimental results show that the improved correction model can be applied to complex weather conditions.(3) Design and implement of corrected data publishing platform. The platform can publish the collected data and the corrected data at the same time. The platform also provides the management of sensor nodes. User can use data query, and other functions. At present, the platform has made a certain social benefits.