Using Big Data Computation for Identifying Unknown Pair Trawlers / 以巨量資料運算方法偵測未知之雙船拖網之船組

碩士 / 國立臺灣海洋大學 / 資訊工程學系 / 107 / The art of fisheries has evolved from small-scale artisanal coastal fishing to a much larger scale industrial fishing, with the revenue produced growing each year. During 2017, fisheries contribute up to 900 billion NTD in economics, with 171 billion belonging to the coastal and offshore fishery in Taiwan. Each year, approximately 180 thousand tons of fish are caught by coastal and offshore fishing vessels. Taiwanese seine catches approximately 85 thousand tons and trawlers catching 40 thousand tons.

During marine resource management, the catch per unit effort (CPUE) is an indirect measurement of the abundance of fishes. However, for pair trawlers, the amount of catch is usually reported from one vessel, making the other vessel invisible in estimation. The result is we have spent double efforts on the catch but calculated it as a single effort, which results in overestimations of the CPUE.

This research uses the voyage data recorder (VDR), which records the trajectories of fishing vessels, as the basis for finding vessel pairs which operate together. Methods provided in this research analyzes and finds potential operation pair trawlers by using the distance and speed information provided. Using the data sets from March 2017, 317 operating pair trawlers were identified with 306 of them being confirmed, constituting an accuracy over 96%. A web interface based on WebGL 3D visualization is also provided that can be used to verify if the vessels are indeed operating in pairs visually. Methods from this research not only can be used in the future for CPUE adjustments for vessels which operate in pairs, but can further be modified to find vessels that operate in teams of three or more.

Identiferoai:union.ndltd.org:TW/107NTOU5394044
Date January 2019
CreatorsTsai, Hsin-Yu, 蔡欣妤
ContributorsWilliam W.Y. Hsu, 許為元
Source SetsNational Digital Library of Theses and Dissertations in Taiwan
Languagezh-TW
Detected LanguageEnglish
Type學位論文 ; thesis
Format38

Page generated in 0.0017 seconds