簡易魚苗自動點算系統設計
國立中山大學海洋生物研究所碩士論文摘要
研究生：黃建華
指導教授：方新疇博士

本研究的目的是研發設計一種簡易魚苗自動點算系統，利用電腦影像處理配合電腦程式進行自動點算，以期達到快速省時及正確的點算結果，可做為養殖漁業上方便使用且低成本的有效工具。

使用像素的色階資料做為指標，辨別出一尾魚苗所佔平均像素後，再利用所有魚苗的面積推算魚苗數，避開傳統點算系統需辨識出個別的目標後才加以計數的方法。本系統使用的點算方法程式演算容易且易於操作。硬體部分包括電腦影像擷取、設備組裝及測試等；軟體部分包括點算程式開發(利用True Basic語言撰寫)，系統操作程序設計及最佳化等。本系統經實際操作後，發現仍然具有偏誤差存在，改善之道則可利用統計方法求得誤差與期望值的迴歸關係予以校正。

在以日光燈魚苗（Paracheirodon innesi）為實驗材料的實際操作中，在60公升水體（49×44×25cm）中放入100、300、500、700及900尾五組不同魚苗數，經本系統點算後的結果，平均值各為132.30、258.75、495.26、798.79及1053.89。其中500尾組(標準差為 88) 的結果表現最佳。全部誤差率在0.9%至32.3%間。

就目前的結果而論，本系統的點算目標之體形大小需近似，且僅適用於本研究中之操作方式 (25cm的水深、點算魚種需為日光燈魚苗等) 。在應用到其它種類的魚苗時須從新定出誤差校正之迴歸方程式。但在使用時，理論上可不受點算目標外觀及水槽形狀大小影響。雖然經由統計方法校正之後誤差值仍大，但和其它目前已知的電腦點算系統比較，其點算結果準確度大多較佳，在操作方便及成本低廉上亦有過之。

A simple design of automatic counting system for fish larvae
Chien-hua Huang
Advisor: Dr. Sun-chio Fong
Institute of Marine Biology, National Sun Yat-sen University, Kaohsiung 804,Taiwan, R.O.C.

The purpose of this study is to develop a portable-computer system for automatic counting of fish larvae. It was intended as an economical, fast and accurate tool for the general fish farmers.

The basic principle of the system is to use gray level value of the three prime colors (red, green and blue) as an indicator to separate and estimate fish larvae within the CCTV image. The hardware part includes a CCTV camera which hooks up to an IBM compatible laptop computer. Images of fish in a tank were captured and stored for future counting process. Program was written in True Basic language for taking the average number of pixel for individual fish, and system optimization for estimating the total number of fishes within tank. Regression analytical methods were also employed for estimating and correction of bias and errors.

In an experiment using the present system to estimate known numbers (100, 300, 500, 700 and 900) of larval Paracheirodon innesi in a 60-liter tank as the testing material, the estimated fish numbers were 132, 259, 495, 799 and 1054 respectively. The group of 500 fishes (standard deviation equals to 88) gave the best result. The total percentage of error ranged between 0.9% and 32%.

Application of this system on other species of fishes is yet to be tested. It was safe to suggest that the current version of the system works only onto the same species of fishes under similar conditions (less than 1000 fishes of similar sizes, water depth of 25 cm etc.). For different target fishes under different type of container, a new set of error-correcting formula would be needed. However, the present results when compared with published reports of other system, are no less accurate. This system however is more economical and works more conveniently than do most others.

目 錄
章 次 頁 數
中文摘要……………………………………………………………….II
英文摘要……………………………………………………………….IV
謝辭…………………………………………………………………….VII
目錄…………………………………………………………………...VIII
表目錄……………………………………………………………...…...IX
圖目錄……………………………………………………………….…..X
壹、 前言………………………………………………………………1
貳、 材料與方法………………………………………………………5
參、 結果………………………………………………………………15
肆、 討論………………………………………………………………19
參考文獻………………………………………………………………..27
表………………………………………………………………………..30
圖………………………………………………………………………..34
附錄……………………………………………………………………..46
個人履歷………………………………………………………………..52

參考文獻
台灣微軟，1998年。 Microsoft Windows98第二版中文使用手冊。126頁。

方新疇，2000年。簡易電腦自動點算魚苗系統設計。農委會計劃書，編號89-自動化-5.1-01(11)。

李振輝，李仁和，1995年。探索影像檔案的奧妙。松崗，台北市。

翁平勝，于錫亮，歐慶賢，2002年。攝影機監控海水槽實驗殘餌率估計。台灣水產學會刊。29(1), 67-71頁。

集可企業，1998年。PhotoMicroGraph Digitize integrate System MGDS 420/EOS。台北。1頁。

登昌恆興業, 1998年。UPMOST UPG400A影像擷取卡中文使用手冊。登昌恆興業股份有限公司，台北。47頁。

蔡玉慧，王蜀嘉，1999年。以色彩資訊協助影像特微之萃取。航測及遙測學刊。第四卷第二期，1-14頁。

Bates, M. C. and Tiersch, T. R., 1997. Low-cost computer-assisted image analysis for fisheries research. The Progressive Fish-Culturist. 59, 235-240.

Chatain, B., Debas, L. and Bourdillon, A., 1996. A photographic larval fish counting technique: comparison with other methods, statistical appraisal of he procedure and practical use. Aquaculture. 141, 83-96.

Foster, M., Petrell, R., Ito, M. R. and Ward, R., 1995. Detection and counting of uneaten food pellets in a sea cage using image analysis. Aqacultural Engineering. 14(3), 251-269.

Hanrahan, B. and Juanes, F., 2001. Estimating the number of fish in Atlantic bluefin tuna (Thunnus thynnus thynnus) schools using models derived from captive school observations. Fishery Bulletin. 99(3), 420-431.

Herman, A. W. and Mitchell, M. R., 1981. Counting and identifying copepod species with an in situ electronic zooplankton counter. Deep-Sea Research. 28A(7), 739-755.

Irvine, J. R., Ward, B. R., Teti, P. A. and Cousens, N. B. F., 1991. Evaluation of a method to count and measure live salmonids in the field with a video camera and computer. North American Journal of Fisheries Management. 11, 20-26.

Jeffries, H. P., Berman, M. S., Poularikas, A. D., Katsinis, Melas, C., I., Sherman, K. and Bivins, L., 1984. Automated sizing, counting and identification of zooplankton by pattern recognition. Marine Biology. 78, 329-334.

Kurts, T. E., 2000. Silver edition reference guide for the True BASIC language system. West Lebanon. NH USA. 642 pp.

Petrell, R. J., Neufeld, T. P. and Savage, C. R., 1993. A video method for noninvasively counting fish in sea cages. Techniques for Modern Aquaculture, American Society of Agricultural Engineers. 352-361.

Ricker, W. E., 1975. Appendix IV. Regression lines. Computation and Interpretation of Biological Statistics of Fish Populations. Information Canadian. Ottawa. 351-352.

Shopov, A., Willams, S. C. and Verity, P. G., 2000. Improvements in image analysis and fluorescence microscopy to discriminate and enumerate bacteria and viruses in aquatic samples. Aquatic Microbial Ecology. 22, 103-110.

Stokes, M. D. and Deane, G. B., 1999. A new optical instrument for the study of bubbles at high void fractions within breaking waves. IEEE Journal of Oceanic Engineering. 24(3), 300-311.

Tsuji, T. and Nishikawa, T., 1984. Automated identification of red tide phytoplankton Prorocentrum triestinum in coastal areas by image analysis. Journal of the Oceanographical Society of Japan. 40, 425-431.

Yang, C. Y. and Chou, J. J., 2000. Classification of rotifers with machine vision by shape moment invariants. Aquacultural Engineering. 24, 33-57.