Using red-green-blue and multispectral camera imagery for stream periphyton monitoring

Session: Poster session

Anika Kuczynski, National Institute of Water and Atmospheric Research, [email protected]
Cathy Kilroy, National Institute of Water and Atmospheric Research, [email protected]
Jochen Bind, National Institute of Water and Atmospheric Research, [email protected]

Abstract

Periphyton plays a vital role in streams, but nuisance biomass levels can cause a variety of health problems, especially in the case of toxic species such as Phormidium, which can be lethal for dogs when ingested. New Zealand’s National Policy Statement for Freshwater Management mandates monthly monitoring of chlorophyll a, the chosen metric for periphyton abundance in streams. Chlorophyll a sampling and visually estimating cover are time- and cost-intensive monitoring methods with low spatial and temporal resolution. Our objective is to test time- and cost-effective imagery-based monitoring, which provides a higher sampling frequency (daily) and more objective results through image analysis algorithms rather than manual sampling with subjective variation. We obtained a time series of daily red-green-blue (RGB) and multispectral (VIS and NIR) imagery of a stream bed in Canterbury, New Zealand. Additionally, we collected weekly ground-truth data of chlorophyll a as well as visual estimates of coverage divided into different groups of periphyton. Although stream characteristics and weather conditions are limiting factors, preliminary image analysis results suggest that both RGB and multispectral imagery can be used for high resolution periphyton monitoring in shallow streams.

1. Keyword
periphyton

2. Keyword
remote sensing

3. Keyword
monitoring

4. Additional Keyword
streams

5. Additional Keyword
multispectral imagery

6. Additional Keyword
New Zealand