The Detection And Tracking Of Fugu Rubripes Based Computer Vision

Rubripes are widely distributed in China, it’s delicious, fresh meat, nutrient-rich, known as the reputation of "a hundred fish of the first", the favorite of consumers, rubripes is high nutritional and economic value of the species one, in improving people’s livelihood and guide people to get rich, the puffer fish farming industry is playing an increasingly important role.This paper study and research image segmentation algorithm and target tracking algorithm about Fugu rubripes based on computer vision.(1)This paper summarizes the history and recent development of applications of computer vision development, but also on fish behavior research methods were reviewed. And of puffer farming methods and models it is described emphatically and pointed out that the research methods of puffer stage mainly through ecological model, that is based on the total input(fed diets) and output(tetrodotoxin production) to study, so the conclusion is only a general concept, the relationship between the individual and the puffer between individuals and between groups cannot be in-depth study of puffer individual and the environment. To some extent, this restricts the development of the puffer fine massive and intensive farming.(2)In this paper, the application of computer vision technology to target identification and target behavior puffer monitoring. This is the first use of computer vision technology to explore and try puffer target identification and target behavior monitoring for computer vision technology to provide a reference and feasibility borrowed from the laboratory to the production plant.(3)In real breeding workshop captures still images, by analyzing the image of the target-rubripes, find out the best features of gray value, gradient features and neighborhood characteristics, the above various features for target feature extraction, and ant colony algorithm rubripes target segmentation. This article farming shop floor actual original image acquisition related research and discussion rubripes target segmentation. Experiments show that this paper has a high accuracy of image segmentation algorithm based on ant colony algorithm.(4)By experimental platform acquisition rubripes motion video, to study the robustness of the tracking algorithm in an aqueous environment, in particular due to the problem of water environment caused by light, water issues and developments background issues and rubripes own velocity fast, sports get together and so proposed tracking algorithm based on Mean-Shift algorithm puffer goal, and to show the process of tracking through a series of pictures. Experimental results show that the proposed tracking algorithm based on tetrodotoxin MeanShift algorithm has better robustness.