节点文献
基于卷积神经网络的植物叶片树种识别研究与实现
Research and Realization of Wood SpeciesRecognition Based on Convolutional Neural Network
【摘要】 随着人工智能的迅速发展,深度学习方向的算法性能逐渐提高,推动了深度学习在各个领域的应用。本文使用卷积神经网络算法建立树种识别模型,以叶片作为模型输入数据。本文所建立的模型在公开的Flavia数据集中的识别准确率在90%以上,达到了应用要求,本模型的设计对林学有一定的实际应用价值。
【Abstract】 As the field of artificial intelligence develop rapidly,the performance of deep learning algorithm is constantly improved,the application of deep learning in various fields is promoted greatly. In this paper,convolution neural network algorithm is used to establish tree species identification model,and leaves are used as model input data. The model established in this paper Flavia The recognition accuracy of data set is 90% Above,the application requirements are met. The design of this model has certain practical application value to forestry.
【关键词】 深度学习;
卷积神经网络;
树种识别;
【Key words】 Deep learning; Convolutional neural network; Wood species recognition;
【Key words】 Deep learning; Convolutional neural network; Wood species recognition;
【基金】 2019年度东北林业大学省级创新项目(201910225226,SJGY20170145)
- 【文献出处】 智能计算机与应用 ,Intelligent Computer and Applications , 编辑部邮箱 ,2020年10期
- 【分类号】TP391.41;TP183;S718.4
- 【被引频次】4
- 【下载频次】237