基于卷积神经网络的养殖鱼类品种识别模型

随着人工智能、大数据、机器学习、计算机视觉等技术的发展，卷积神经网络（CNN）越来越多地应用于图像识别领域，图像数据集的丰富性以及多样性对CNN模型的性能和表达能力至关重要，但现有的鱼类图像公共数据集资源较匮乏，严重缺少训练集以及测试集样本，难以满足深度CNN模型优化及性能提升的需要。实验以大黄鱼、鲤、鲢、秋刀鱼和鳙为对象，采用网络爬虫以及实验室人工拍照采集相结合的方式，构建了供鱼种分类的基础图片数据集，针对网络爬虫手段获取到的鱼类图像存在尺度不一、格式不定等问题，采用图像批处理的方式对所有获取到的图像进行了统一的数据预处理，并通过内容变换以及尺度变换对基础数据集做了数据增强处理，完成了7 993个样本的图像采集与归纳；在权值共享和局部连接的基础上，构建了一个用于鱼类识别的CNN模型，采用ReLU函数作为激活函数，通过dropout和正则化等方法避免过度拟合。结果显示，所构建的CNN鱼种识别模型具有良好的识别精度和泛化能力。随着迭代次数的增加，CNN模型的性能也逐步提高，迭代1 000次达到最佳，模型的准确率为96.56％。该模型采用监督学习的机器学习方式，基于CNN模型，实现了5种常见鱼类的鱼种分类，具有较高的识别精度和良好的稳定性，并且模型本身具有一定程度的鱼种推广可能性，为养殖鱼类的品种识别提供了一种新的理论计算模型。

Recognition model of farmed fish species based on convolutional neural network

With the development of artificial intelligence, big data, machine learning, computer vision and other technologies, convolutional neural networks (CNN) were increasingly used in the field of image recognition, which greatly improved the efficiency and accuracy of recognition. The richness and diversity of image data sets were crucial to the performance and expressive ability of convolutional neural network models. However, the existing fish image data set resources are relatively scarce, and the training set and test set samples are severely lacking, and it is difficult to meet the needs of deep convolutional neural network model optimization and performance improvement. A basic image data set for fish species classification was constructed by using a combination of web crawlers and manual camera collection in the laboratory. Large yellow croaker, silver carp, catfish, saury and bighead carp were used as the test objects in this paper. In view of the problems of different scales and uncertain formats of images, image batch processing, unified data preprocessing was performed on all the acquired images, and the basic data set was enhanced through content transformation and scale transformation, and the image collection and induction of 7 993 samples were completed. On the basis of parameter sharing and local connectivity, a convolutional neural network model for fish recognition were constructed; the ReLU function was used as the activation function to improve the performance of the algorithm; the dropout and regularization were used to avoid overfitting. The test results showed that: the convolutional neural network fish species recognition model constructed in this paper have good recognition accuracy and generalization ability. As the number of iterations increases, the performance of the convolutional neural network model gradually improves. It reached to best when the number of iterations came to 1 000 times. The accuracy of the model was 96.56%. The model adopted the machine learning method of supervised learning. Based on the CNN model, it realized the classification of five common fish species, with high recognition accuracy and good stability. The model itself had a certain degree of possibility of fish species promotion and rovided a new theoretical calculation model for the species identification of farmed fish.