Recognition and Classification of Ornamental Fish Image Based on Machine Vision

Abstract

With the gradual maturity of smart home technology, it requires a lot of knowledge and experience to judge the life state of ornamental fish artificially when raising them in a traditional aquarium in a smart home environment. Obviously not everyone can accurately judge the life state of ornamental fish. Accurate judgment, therefore, ornamental fish life state and timely feedback is essential for pet fish. Aiming at the actual need of intelligent feeding of traditional ornamental fish, an application of image recognition and classification of ornamental fish based on machine vision was designed. The fish classification model based on Convolutional Neural Network (CNN) is studied, and based on the model, the training of ornamental fish classification model is further realized by combining migration learning. The experiment used TensorFlow to train the network model. The experimental results show that the recognition accuracy of ornamental fish can reach 98.1% by using this method, and the corresponding knowledge base and rule base are constructed by combining with the expert system, which effectively solves the problem of insufficient knowledge and experience in artificial feeding of ornamental fish.