摘要
针对高分对地观测系统使用过程中会受到不同活动项目的约束影响,出现系统成像、回传及活动完成率低的问题,导致观测效果不佳,为此提出了基于尺度特征卷积神经网络的高分对地观测系统设计;该系统通过管控中心服务器推送系统运行状态信息,实现三维显示任务的功能;利用CMOS图像传感器实现成像面对应点的传送和FPGA控制器控制其数据存储时间;采用BCM5464千兆交换机,实现数据高速传输;构建并训练尺度特征卷积神经网络,利用RPN网络识别目标区域特征,通过划分目标的前景和背景确定了该区域内的训练兴趣区域坐标,从而使RPN网络权值学习达到了预期目标,提升了目标检测识别的准确性,设计对地观测信息管理流程,完成系统设计;由实验结果可知,该系统最高成像、回传概率、活动完成率分别为83%、99.9%和100%,具有良好的观测效果。
In view of the high-scoring earth observation system being affected by the constraints of different activities during the use process,the system imaging,backhaul and activity completion rate is low,resulting in poor observation results,for this reason,a convolutional neural network based on scale features is proposed.The design of the high-scoring Earth observation system.The system pushes system operating status information through the management and control center server to achieve the function of three-dimensional display tasks.The CMOS image sensor is used to realize the transmission of the corresponding points on the imaging surface,and the FPGA controller is used to control the data storage time.Adopt BCM5464 gigabit switch to realize high-speed data transmission.Construct and train the scale feature convolutional neural network,use the RPN network to identify the characteristics of the target area,and determine the training interest area coordinates in the area by dividing the foreground and background of the target,so that the RPN network weight learning achieves the expected goal and improves The accuracy of target detection and recognition,the design of the Earth observation information management process,and the completion of the system design.It can be seen from the experimental results that the highest imaging,return probability,and activity completion rate of the system are 83%,99.9%,and 100%,respectively,which has good observation effects.
作者
刘笛
何伟
曹秀云
LIU Di;HE Wei;CAO Xiuyun(China Aerospace Academy of Systems Science and Engineering,Beijing 100045,China)
出处
《计算机测量与控制》
2021年第12期215-219,共5页
Computer Measurement &Control
关键词
尺度特征
卷积神经网络
高分对地观测
图像传感器
scale feature
convolutional neural network
high-resolution earth observation
image sensor
