摘要
为了应对动物保护工作者和普通民众希望对国内珍稀动植物准确识别并加以保护的场景,实验提出一种基于数据增强技术和迁移学习的深度学习方法,并借助App实现实时的珍稀动植物识别。首先使用迁移模型提取基于ImageN.et数据集的标准化瓶颈描述算子和所有卷积层神经网络的权重,然后通过网络爬虫搜集大量野生珍稀动植物图片,并对得到的数据集分别进行静态数据增强和动态数据增强,用来训练迁移模型新的特征表示。实验使用Softmax函数实现多物种分类。实验中,分别采用MobileNet、InpectionV3等不同模型的不同参数进行训练,得到不同配置下模型的准确率并加以比较,最后将其应用于真实自然场景,实现了90%以上的珍稀动植物精准分类,实验证明提出的方法具有很高准确率和良好的运行性能。
In order to deal with the scenes where animal protection workers and ordinary people want to accurately identify and protect rare plants and animals in China,proposes a deep learning method based on data enhancement technology and migration learning,and realizes real-time rare animal and plant identification by means of App.Firstly,the migration model is used to extract the weight of the standardized bottleneck description operator and all bottleneck layer activation functions based on the ImageNet dataset.Then,a large number of wild rare animal and plant images are collected through the web crawler,and the data set is separately subjected to static data enhancement and dynamic data enhancement.Used to train the new feature representation of the migration model.The experiment uses the Softmax function to achieve multi-species classification.In the experiment,different parameters of different models such as MobileNet and InpectionV3 are used to train the test,and the accuracy of the model under different configurations is tested.Finally,applies to the actual natural scene,realizes more than 90%of the rare and accurate real-time classification of rare animals and plants.The proposed method has high accuracy and good running performance.
作者
宋益盛
林志杰
SONG Yi-sheng;LIN Zhi-jie(School of Electric and Information,Shanghai Dianji University,Shanghai 100120)
出处
《现代计算机》
2019年第14期57-63,共7页
Modern Computer
基金
大学生科创项目(No.A1-0224-18-012-073)
作者简介
宋益盛(1996-),本科,研究方向为人工智能在图像识别领域的应用;林志杰(1980-),女,博士,项目指导教师,研究领域为大数据分析与挖掘.
