Design And Implement Of The Method Of Constructing Noise MAP Based On Crowed Sensing

Noise pollution becomes more severe compared with other type of pollution,and it influences humans life in many aspects.However,there are few approaches for people realize where the noise comes from and what makes some regions so noisy.Government is incapable of controlling noise pollution due to lack of the distribution of urban noise.Humans will still suffer from noise pollution if no action is taken to sense and eliminate the urban noise.The contribution of this paper is to construct noise map using mobile crowd sensing.The core idea is using ”humans as sensor” to sense urban noise everywhere.Firstly,we develop an APP in smartphone equipped with microphone to detect environment noise.Since the microphone can not perform as well as professional noise meter,we discover the relationship between noise data collected by smartphone and noise meter based on large noise sample,then we propose a method to calibrate noise data of smartphone to make it more accurate.Secondly,the main work of this paper is to recover missing noise data based on sparse sensing value.We propose noise tensor model and recover missing entries by tensor completion.Combining with high correlation features,we proposed semi-supervised tensor completion to improve recovery accuracy.Finally,we construct Shanghai noise map.