深度学习在林业中的应用  

Application of Deep Learning to Forestry

作　　者：南玉龙 张慧春[2] 郑加强[2] 杨琨琪 Nan Yulong;Zhang Huichun;Zheng Jiaqiang;Yang Kunqi(College of Mechanical Engineering,Yancheng Institute of Technology,Yancheng 224051,Jiangsu,China;College of Mechanical and Electronic Engineering,Nanjing Forestry University,Nanjing 210037,China)

机构地区：[1]盐城工学院机械工程学院,江苏盐城224051 [2]南京林业大学机械电子工程学院,南京210037

出　　处：《世界林业研究》2021年第5期87-90,共4页World Forestry Research

基　　金：盐城工学院校级科研项目资助(xjr2021012);江苏省现代农机装备与技术示范推广项目(NJ2020-18);国家自然科学基金项目(32171790);江苏省六大人才高峰(NY-058);江苏省青蓝工程项目(苏教201842)。

摘　　要：林业数据信息的智能处理与深度挖掘分析是智慧林业精准决策与统筹管理的基础。文中主要从树木识别与分类、森林火灾识别与预测、树木病虫害监测和木材检测等4个方面总结深度学习在林业中的应用;从林业数据集、数据标记、特征学习涵盖范围、图像遮挡等方面分析了深度学习在林业领域应用上的缺点与局限性,以充分认识深度学习在林业应用方面亟待解决的问题与突破的方向;从林业领域应用场景、算法改进、数据集共享与研究成果应用的角度,展望深度学习在林业领域应用上亟待加强研究的方面,以提升深度学习解决林业应用问题的广度与深度,促进智能林业的发展。The intelligent processing and deep mining analysis of forestry data information are the basis for intelligent and precise forestry decision-makings and integrated management. This paper summarizes the application of deep learning to forestry from four aspects of tree identification and classification, forest fire identification and prediction, tree pest and disease monitoring and wood detection, and analyzes the shortcomings and limitations of deep learning application in forestry in terms of forestry datasets, data labeling, feature learning coverage, and image occlusion, so as to fully understand the problems and breakthrough directions of deep learning application in forestry. From the perspectives of application scenarios in forestry, algorithm improvement, dataset sharing and research results application, the future researches on the application of deep learning to forestry that need to be strengthened are prospected, in order to enhance the breadth and depth to deep learning application in forestry and promote the development of intelligent forestry.

关 键 词：深度学习 树木识别 森林火灾预测 病虫害监测 木材检测 

分 类 号：S7[农业科学—林学] TP391[自动化与计算机技术—计算机应用技术]