This study applies Unmanned Aerial Vehicle (UAV) to netcage fish-farming. The UAV can fly thru cages along a predefined route. At each cage, the UAV can automatically drop sensors to collect the environment data around the cage. However, low cost commercial or assembled drone might not have precise Global Navigation Satellite System (GNSS). Also, there are many uncertain factors cause Global Positional System (GPS) instability, like weather, location…etc. The main idea of this study is using GPS to guide the drone to approximated location of each cage. Then, apply image recognition to obtain the net-cage and UAV relative position. After that, the drone can utilize this information to adjust its position to the desired target. In this study we use fully assembled drone that is controlled by the Pixhawk, and the main processing platform is the Raspberry Pi3.
Chen, Chao-Xun and Juang, Jih-Gau
"Vision Based Target Recognition for Cage Aquaculture Detection,"
Journal of Marine Science and Technology: Vol. 28
, Article 2.
Available at: https://jmstt.ntou.edu.tw/journal/vol28/iss6/2