Share Email Print

Proceedings Paper

BLUEWATER EYE: using satellite as a low cost water pollution sensor: analytics for deriving long term pollution insights based on mapping water turbidity
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Bluewater EYE analytics bring the power of remote sensing backed by artificial intelligence to address several water pollution related challenges at low cost as it requires minimal calibration with Internet of Things (IoT) data and can provide useful water quality insights via spatio-temporal pollution signatures mapped over possibly large distances, applicable to any global water body. We leverage geo-spatial data analytics in combination with cloud based, real-time in-situ sensing, that is capable of mapping changes in water turbidity of a water resource. Spatio-temporal pollution insights are derived and represented via color-based heatmaps and matrices that are overlaid on Google maps and facilitate visualization and identification of pollution hotspots. Methods include first-order correlations derived from surface reflectance of different satellite band/s (single or a normalized difference combination) and high resolution, geo-tagged, in-situ turbidity sensing via a moving multi-parameter sensor platform. Significant improvements in correlation (upto 80%) were obtained by using statistical methods such as moving average for filtering out the sensing data associated noise. Moreover, we use machine learning based approach for training a turbidity model based on Support vector machine (SVM) regression with Radial Basis Kernel (RBF). The predicted turbidity values for the selected region based on Landsat 8 Level 2 surface reflectance data, was then applied for time-series data for historical years of 2016 and 2017. The turbidity time series so obtained is analyzed to capture any significant variations in water quality due to various factors (event or season based). We propose a method to analyze pollution contributions from different individual sources, that is based on assigning an individual spatio-temporal signature in the form of a color-coded matrix. Overall, Bluewater EYE could be useful for capturing water quality spatio-temporal information and trends, providing data driven proofs for timely alerts and advisory and other useful insights for implementation of appropriate remediation measures and interventions– all this at an affordable cost.

Paper Details

Date Published: 10 October 2018
PDF: 13 pages
Proc. SPIE 10783, Remote Sensing for Agriculture, Ecosystems, and Hydrology XX, 107831D (10 October 2018); doi: 10.1117/12.2325589
Show Author Affiliations
Sukanya Randhawa, IBM Research - India (India)
Ranjini B. Guruprasad, IBM Research - India (India)
Srinivas Rao Balivada, Univ. of Chicago Ctr. in Delhi (India)
Priyank Hirani, Univ. of Chicago Ctr. in Delhi (India)
Supratik Guha, The Univ. of Chicago (United States)

Published in SPIE Proceedings Vol. 10783:
Remote Sensing for Agriculture, Ecosystems, and Hydrology XX
Christopher M. U. Neale; Antonino Maltese, Editor(s)

© SPIE. Terms of Use
Back to Top