您的位置:首页  > 论文页面

数据缺失情况下基于压缩感知的农业传感数据采集与还原方法

发表时间:2016-08-15  浏览量:1222  下载量:487
全部作者: 刘 峰
作者单位: 华中农业大学信息学院
摘 要: 在农业物联网中,由于传感节点能量受限,故需要设计高效的数据采集算法,同时传感节点所处环境恶劣常会导致传感数据失效。针对此问题,基于压缩感知理论设计并实现了数据缺失情况下的农业传感数据采集与还原方法。首先,采用随机矩阵对农业传感数据进行压缩采集,使采集的数据样本数远低于实际的传感数据个数,可降低传感器的能耗。其次,采用1范数最优的方法从采集的样本中还原出原始数据。同时为了反映农业传感数据的缺失情况,在测试数据集中将一些连续的原始数据设置为0,以表示数据的缺失。结果表明,该方法能够还原原始传感数据,同时采集样本数的增加可降低数据还原的误差,但当采集样本数目达到一定程度时误差降低较为缓慢。
关 键 词: 农业机械化工程;传感数据压缩采集;传感数据还原方法;压缩感知
Title: Agricultural sensor data collection and recovery method based on compressed sensing under the condition of data missing
Author: LIU Feng
Organization: College of Informatics, Huazhong Agricultural University
Abstract: In agricultural internet of things, because of the limitation of sensor’s energy, some efficient data collection methods should be designed and the hostile environments where the agricultural sensors are deployed often cause the failure of sensor data. To solve these problems, an agricultural sensor data collection and recovery method, which takes the sensor data failure into consideration, is proposed based on compressed sensing theory. In this method, firstly a random sample collecting matrix was used to collect the sensor data, which can collect much less samples than original sensor data and reduce the consumption of energy. Secondly, the 1-norm optimization algorithm was employed to recover the original data. In order to simulate the data failure phenomenon, some continuous sensor data were set to zero which indicates the failure of sensor data. The results show that by this method original sensor data can be recovered, and increasing the number of collected samples will reduce the recovery error. But when the number of collected samples exceeds certain threshold, the recovery error reduces slowly.
Key words: agricultural mechanization engineering; sensor data compressive collection; sensor data recovery method; compressed sensing
发表期数: 2016年8月第15期
引用格式: 刘 峰. 数据缺失情况下基于压缩感知的农业传感数据采集与还原方法[J]. 中国科技论文在线精品论文,2016,9(15):1553-1557.
 
1 评论数 0
暂无评论
友情链接