Integrating Multisensor Satellite Data Merging And Image Reconstruction In Support Of Machine Learning For Better Water Quality Management

Keywords

Enabling technology; Machine learning; Remote sensing; Water quality; Watershed management

Abstract

Monitoring water quality changes in lakes, reservoirs, estuaries, and coastal waters is critical in response to the needs for sustainable development. This study develops a remote sensing-based multiscale modeling system by integrating multi-sensor satellite data merging and image reconstruction algorithms in support of feature extraction with machine learning leading to automate continuous water quality monitoring in environmentally sensitive regions. This new Earth observation platform, termed “cross-mission data merging and image reconstruction with machine learning” (CDMIM), is capable of merging multiple satellite imageries to provide daily water quality monitoring through a series of image processing, enhancement, reconstruction, and data mining/machine learning techniques. Two existing key algorithms, including Spectral Information Adaptation and Synthesis Scheme (SIASS) and SMart Information Reconstruction (SMIR), are highlighted to support feature extraction and content-based mapping. Whereas SIASS can support various data merging efforts to merge images collected from cross-mission satellite sensors, SMIR can overcome data gaps by reconstructing the information of value-missing pixels due to impacts such as cloud obstruction. Practical implementation of CDMIM was assessed by predicting the water quality over seasons in terms of the concentrations of nutrients and chlorophyll-a, as well as water clarity in Lake Nicaragua, providing synergistic efforts to better monitor the aquatic environment and offer insightful lake watershed management strategies.

Publication Date

10-1-2017

Publication Title

Journal of Environmental Management

Volume

201

Number of Pages

227-240

Document Type

Article

Personal Identifier

scopus

DOI Link

https://doi.org/10.1016/j.jenvman.2017.06.045

Socpus ID

85021280402 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/85021280402

This document is currently not available here.

Share

COinS