The invention discloses a
wireless sensor network
temperature monitoring method based on an improved
Kriging algorithm. The
wireless sensor network
temperature monitoring method comprises the steps of: firstly, establishing a layer
cluster type wireless sensor network topological structure, analyzing historical temperature data of a target region, setting a threshold value rho and a fusion cycle T, performing periodic fusion on in-
cluster node data at a cluster head node based on the threshold value rho, acquiring initial temperature data, transmitting the initial temperature data to a monitoring host by means of a wireless link and saving the initial temperature data in a local
database; importing the initial temperature data, and performing variable
structure analysis on the initial temperature data to generate a variation function expression; and optimizing a variation
function model by utilizing an improved Nelder-Mead simplex method, designing a
regular grid terrain, performing temperature interpolation by adopting an ordinary
Kriging algorithm, carrying out error correction, and outputting a
regular grid temperature
field pattern. The
wireless sensor network temperature monitoring method optimizes the variation
function model by utilizing the improved Nelder-Mead simplex method, corrects weights of the ordinary
Kriging algorithm, performs error compensation on the interpolation result, and makes the interpolation result more scientific and closer to the actual situation.