Classification of Fish Species by Processing the Hydroacoustic Signal and Canonical Discriminant Analysis

Abstract

本研究應用影像處理技術並配合掃描器，開發魚羣量指標之計測系統(EIPS,Echo-tracesImage Processing System)以取代昔日人力測量方式。經實測與統計比較結果，發現使用本套魚羣量計測系統具有準確度高，速度極快，記憶與處理能力極大，以及提高校正與診斷等功能。另外再使用本套魚羣量計測系統，實測鮆科魚羣魚探記錄跡，並與實際漁獲量比較得到相當高之相關性。In order to improve the estimated effectiveness of fish abundance indexfrom echo traces,an echo-traces image processing system(EIPS)was designedto incorporate with digital image processing with a scanner and a PC/AT.A traditional scale-lupe method was used to compare with the EIPS program.As the results,the EIPS program estimation has the advantages of higheraccuracy and speed,large memory and effectiveness in comparison with thetraditional scale-lupe method.Further,the EIPS program was applied to the estimation of six acousticsurveys for anchovy larvae.The estimates of abundance index were obviouslyrelated to actual catches.