Please use this identifier to cite or link to this item:
Development of an underwater video system for white shrimp (Litopenaeus vannamei) in ponds
|Authors: ||NG WOEI LING|
|Contributors: ||NTOU:Department of Aquaculture|
underwater video system;turbidity;shrimp recognition;feeding management
|Issue Date: ||2020-07-03T08:29:55Z
|Abstract: ||本研究旨在開發白蝦養殖池的水下影像系統，監測一般混濁低透明度水域蝦池中的白蝦活動，即時取得蝦子大小及進食活動等資訊，提供人工智慧控制投餌最佳化。目前國內外業者礙於蝦池懸浮顆粒與透明度等問題，仍無法克服影像清晰度的瓶頸，本研究得知在低透明度的養殖水中，紅外線比白光可拍攝更遠距離的清晰影像。本研究設計三種影像設備，利用透明水箱分隔水下攝影機和蝦子，可以大幅增加清晰影像的範圍。影像設備I為玻璃材質所組成的透明四方体；影像設備II為一梯形構造，底部設有一觀察平台；影像設備III水箱同為梯台型，水箱上端為透明觀察平台。影像設備I與影像設備III在透明度10cm的養殖水中取得可辨識白蝦的影像;而影像設備II的取得清晰影像的極限為透明度30cm。 本研究改良影像設備I透明水箱的大小，以距水前20、 30公分所得影像效果最佳，拍攝角度與影像畫面更為寬闊。但影像設備I無法清楚取得水下飼料的影像，但可取得餵食一小時內，出現白蝦影像的數目，以反映其攝食強度。影像設備III可同時清楚觀察蝦子攝食的行為和飼料殘留量，但觀察平台的高度會影響白蝦到設備上攝食的意願；在設備上投放1%池中飼料投餵量與傳統傘網的殘餌結果相近。以上結果顯示，在影像設備III取得之清晰影像可用於調整飼料投餵量，方法為投放1%的飼料量在設備平台上，觀察投餵1小時後的飼料殘留變化，再依此調整下一次的投餵量。 利用影像設備I與影像設備III所拍攝的水下白蝦影像，可透過電腦分析標示出可辨識的白蝦並估算蝦體長度，有助於建立智慧化的蝦池管理系統。本研究所建立的水下影像系統已克服混濁水域影像清晰度的問題，提供投餌控制與白蝦健康狀況訊息，可減少飼料的浪費與加強養殖管理，未來可朝向自動投餵與數據庫建立的方向發展。|
The purpose of this study is to develop an underwater video system for white shrimp culture ponds, monitoring shrimp size and feeding behavior in unclear and low-transparent water in shrimp ponds which provides the optimization of feed control with automated artificial intelligence. The previous technique was difficult and impossible to have a clear image of the shrimp when the water in ponds were turbid and low-transparent. Therefore, IR (Infrared) light was designed to capture more clear images for low trasparence water. Three types of video devices were constructed by using a transparent water tank to separate underwater cameras and shrimps, which can greatly increase the wide scope of clear images. Device (I) is made by glass while capturing device (II) and (III) are trapezoidal structure with the bottom of transparent observation platform. Device (I) and (III) have a good performance in the water with transparency 10cm, while the limit of device (II) achieves transparency 30cm. The video device (I) is improved the size of the water tank to 20-30cm, showed the greatest performances with a wider range of view. However, device (I) can capture the number of white shrimps present within one hour of feeding which was calculated to reflect their feeding intensity but can’t capture properly of the feed pellets in the video. According to the results of the experiment, the height of the device (III) affects the willingness of the white shrimp to consume feed but it can clearly observe the shrimps and uneaten feed. The results showed that, the residual feed of 1% feed on the device was the closest to the traditional feed tray. Device (III) provide farmers a clearer image to adjust the feed amount by placing 1% feed on the device platform, then observe the change of feed residue after 1 hour of feeding, and optimize next feed amount. The white shrimp imaging taken by device (I) and (III) can estimate the length of the shrimp, which will help to establish an intelligent shrimp pond management system. The establishment of the underwater video system has solved the problem of image clarity in turbid waters, providing information on feed management and health status of white shrimp, making a possible immediate response on overfeeding and strengthen aquaculture management., This study can be developed towards automatic feeding and database establishment in the future.
|Appears in Collections:||[水產養殖學系] 博碩士論文|
All items in NTOUR are protected by copyright, with all rights reserved.