碩士 / 國立臺灣大學 / 生物產業機電工程學研究所 / 89 / A monitoring system is developed using machine vision to analyze the fish weight in a re-circulating raceway aquaculture system. Using this system, we can obtain the weight distribution of the fish in culturing ponds to be used for harvesting, sorting, or feeding operations. The system uses a CCD to capture the images of the fish passing through an underwater transparent channel. Because the fish stays in water during image capturing, the stress is close to eliminated. Also because of the characteristics of the raceway aquaculture system, complete sampling is made possible. The study is done using tilapia (Tilapia nilotica). Data was taken in laboratory to determine the relationship area and weight between the area regression equation (the coefficient of determination is 96.38%) has higher accuracy than the weighted area regression equation. To increase the speed of the fish passing through the channel a counter flow is used to guide the fish into the channel. The maximum passing efficiency is 132 fish/hr at 6.7 cm/s. Fish images are taking in the field and brought to the lab for analysis. The image-processing algorithm discards Repetitive and Overlapping images automatically. The image processing time was 4.626 (second/picture). As a result, the effective image was 28% of the total fish passing through the channel. The overall accuracy of the system is 94%. The error comes largely from the opening or shrinking of the fish fins, which causes differences in the fish area. The main problem of the system is due to slow fish movement thought the channel which could extend the sampling time and cause oxygen deficiency in the holding area.
Identifer | oai:union.ndltd.org:TW/089NTU00415003 |
Date | January 2001 |
Creators | Ho, Chien-Chung, 何建忠 |
Contributors | YUAN-NAN CHU, 朱元南 |
Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
Language | zh-TW |
Detected Language | English |
Type | 學位論文 ; thesis |
Format | 0 |
Page generated in 0.0017 seconds