Chapter

AGILE 2015

Part of the series Lecture Notes in Geoinformation and Cartography pp 145-163

Date:

Towards Real-Time Processing of Massive Spatio-temporally Distributed Sensor Data: A Sequential Strategy Based on Kriging

  • Peter LorkowskiAffiliated withInstitute for Applied Photogrammetry and Geoinformatics (IAPG), Jade University of Applied Sciences Wilhelmshaven/Oldenburg/Elsfleth Email author 
  • , Thomas BrinkhoffAffiliated withInstitute for Applied Photogrammetry and Geoinformatics (IAPG), Jade University of Applied Sciences Wilhelmshaven/Oldenburg/Elsfleth

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Sensor data streams are the basis for monitoring systems which infer complex information like the excess of a pollution threshold for a region. Since sensor observations tend to be arbitrarily distributed in space and time, an appropriate interpolation method is necessary. Within geostatistics, kriging represents a powerful and established method, but is computation intensive for large datasets. We propose a method to exploit the advantages of kriging while limiting its computational complexity. Large datasets are divided into sub-models, computed separately and merged again in accordance with their kriging variances. We apply the approach to a synthetic model scenario in order to investigate its quality and performance.

Keywords

Continuous phenomena Sensor data streams Spatio-temporal interpolation Kriging Deviation map