S&M2076 Research Paper of Special Issue
Published: December 26, 2019
Early Detection System of Harmful Algal Bloom Using Drones and Water Sample Image Recognition [PDF]
Fukuyoshi Kimura, Akihiro Morinaga, Masayoshi Fukushima, Tomonari Ishiguro, Yasuhiko Sato, Akihiro Sakaguchi, Tomoyuki Kawashita, Ikuo Yamamoto, and Toru Kobayashi
(Received April 29, 2019; Accepted August 6, 2019)
Keywords: image recognition, harmful algal bloom, drone, web application, convolutional neural network (CNN)
Food consumption is increasing as the world population increases. While the eating of fish is spreading worldwide, the depletion of fishery resources has become a problem owing to overfishing, and the importance of aquaculture is increasing in order to continue to supply fish as food. In marine aquaculture, fish are grown in aquaculture cages in the sea. Thus, if a harmful algal bloom (HAB) reaches the cages, it will cause serious damage. Countermeasures against HAB are one of the important problems for aquaculture fishermen. In Nagasaki Prefecture in Japan, countermeasures against HAB include patrolling and sampling water by ships. These samples are then submitted to HAB experts for analysis. Following this analysis, notification is provided to aquaculture fishermen. When HAB is detected early, aquaculture fishermen can minimize the damage of HAB by stopping feeding and moving the aquaculture cages. In this study, we developed a system of early detection and notification of HAB to aquaculture fishermen. This is carried out by patrolling and sampling water using drones, detection of HAB by microscope and PC operation, and automatic notification by email and web application. As a result of this developed system, the identification accuracy of harmful plankton is more than 90%, and the time taken to find HABs can be shortened from the 6 h of the conventional approach to 15 min using this newly developed system.
Corresponding author: Fukuyoshi Kimura
This work is licensed under a Creative Commons Attribution 4.0 International License.
Cite this article
Fukuyoshi Kimura, Akihiro Morinaga, Masayoshi Fukushima, Tomonari Ishiguro, Yasuhiko Sato, Akihiro Sakaguchi, Tomoyuki Kawashita, Ikuo Yamamoto, and Toru Kobayashi, Early Detection System of Harmful Algal Bloom Using Drones and Water Sample Image Recognition, Sens. Mater., Vol. 31, No. 12, 2019, p. 4155-4171.