Context and objectivesKnowledge about the distribution of zooplankton is a key factor in predicting accurately the population dynamics of commercially important fish species and in establishing the potential of sustainable aquaculture operations. Most related application developments have been concentrated on studying phytoplankton in the seas. The sea water colour processing has focused evidently on the greenish colour spectrum, generated by chlorofyl. Zooplankton, instead, is characterized by its general lack of chlorofyl and the more common reddish colour spectrum generated by carotenoid and affiliated compounds. Furthermore, these developments pertained to wide views of the oceans to study this biomass on a macroscale. We, instead, are interested on a derivative of the biomass on a mesoscale. This requires a resolution on the borderline of what most commercially available sensors provide. Also, most developments are geared towards monitoring more global changes within the time frame of several seasons. We are interested in detecting phenomena that are only relevant within a period of days as winds and currents disperse these clouds of organic material. A similar mode of operation is demonstrated by oil spill detection. The main objective of this study is to determine the technical feasibility of a detection system for zooplankton in lakes, based on present or near-future satellite sensors.
Expected scientific results
Only one multispectral SPOT image could be selected that was taken during the collection period of the ground truth data at the Great Salt Lake (USA). This image was used as an input for the different image processing techniques that had the potential possibility to distinguish between the zooplankton streaks and their neighbourhood. None of the presented methods could provide evidence of streaks in the image. Probably, the resolution of the used image wasn’t high enough to create a good filter for the detection of the streaks. Further research is needed to investigate the spectral characteristics of the zooplankton. When these characteristics will be known there is a much higher possibility to find a good algorithm for the detection of the streaks in high resolution imagery.