复杂战场环境的图像训练样本构建技术  

Image Training Sample Construction Technique for Complex Battlefield Environment

作　　者：李航宇 吴云超 李蔚清[1] 苏智勇[3] LI Hangyu;WU Yunchao;LI Weiqing;SU Zhiyong(School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China;Science and Technology on Information System Engineering Laboratory,Nanjing 210023,China;School of Automation,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学计算机科学与工程学院,南京210094 [2]信息系统工程重点实验室,南京210023 [3]南京理工大学自动化学院,南京210094

出　　处：《指挥信息系统与技术》2022年第4期40-45,68,共7页Command Information System and Technology

基　　金：装备发展部“十四五”预研课题(50904040201);军委科技委基础加强计划(2020-JCJQ-ZD-007-00);信息系统工程重点实验室开放基金(05202005)资助项目。

摘　　要：为了解决战场环境图像训练样本种类少、数量小的问题,基于UE4(虚幻4三维渲染引擎)、3DMax和MVC(模型、视图和控制器)模式架构等建立了复杂战场图像训练生成系统。该系统综合了各种影响战场环境图像生成的因子,借助于三维引擎构建了一个虚拟战场仿真环境。该虚拟战场环境可用于仿真图像生成以丰富战场环境样本,从而满足人工智能算法的需要,可防止出现过拟合现象。To solve problems of small variety and little quantity of image training samples in battlefield environment,a complex battlefield image training generation system is established based on UE4(unreal engine 4),3DMAX and MVC(model,view,controller)pattern architecture structure system.Various impact factors on the image generation in battlefield environment are integrated in the system.With the help of 3D engine,a virtual battlefield simulation environment is constructed.The virtual battlefield environment can be used to generate simulation images to enrich battlefield environment samples.Thus,the need of artificial intelligence algorithm can be meeted,and the overfitting phenomenon can be prevented.

关 键 词：虚拟战场 UE4 3D建模 数字高程模型 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]