Polytetrafluoroethylene (Teflon), a widely used spacecraft material, isstudied to investigate the vacuum ultraviolet (VUV) effects and its synergistic effects with atomicoxygen (AO) in a ground-based simulation facili...Polytetrafluoroethylene (Teflon), a widely used spacecraft material, isstudied to investigate the vacuum ultraviolet (VUV) effects and its synergistic effects with atomicoxygen (AO) in a ground-based simulation facility. The samples before and after the experiments arecompared in appearance, mass, optical properties and surface composition. The reactioncharacteristics of Teflon are summarized and the reaction mechanisms are analyzed. The followingconclusion can be drawn: at the action of VUV the Teflon sample surface is darkened for theaccumulation of carbon; and when the sample is exposed to AO, the carbon is oxidized and thedarkening surface is bleached; the synergistic effects of VUV and AO may cause the erosion of Teflonmore severe.展开更多
地面目标伪装效果评估是基于目标在探测频段下伪装前后检测概率的变化,合成孔径雷达能够全天候工作,分辨率较高,已经成为对地侦查的主要手段。对于SAR图像的目标检测概率通常采用专家判读和特征模板匹配等方式获取,这些方法主观性较强...地面目标伪装效果评估是基于目标在探测频段下伪装前后检测概率的变化,合成孔径雷达能够全天候工作,分辨率较高,已经成为对地侦查的主要手段。对于SAR图像的目标检测概率通常采用专家判读和特征模板匹配等方式获取,这些方法主观性较强或对算法设计的依赖度较高,因此需要一种科学客观的评判方法。随着神经网络在图像识别领域的快速发展,目标特征提取和识别结果愈加准确,现采用YOLO-V3(You Only Look Once)算法模型,对MSTAR目标背景合成图像进行机器学习,自动获取目标特征进行识别。训练后的模型在测试集的识别准确度为97.87%,用训练好的模型对未伪装样本与伪装后样本分别进行检测概率计算,通过比较检测概率的变化从而量化某类目标的伪装效果。展开更多
文摘Polytetrafluoroethylene (Teflon), a widely used spacecraft material, isstudied to investigate the vacuum ultraviolet (VUV) effects and its synergistic effects with atomicoxygen (AO) in a ground-based simulation facility. The samples before and after the experiments arecompared in appearance, mass, optical properties and surface composition. The reactioncharacteristics of Teflon are summarized and the reaction mechanisms are analyzed. The followingconclusion can be drawn: at the action of VUV the Teflon sample surface is darkened for theaccumulation of carbon; and when the sample is exposed to AO, the carbon is oxidized and thedarkening surface is bleached; the synergistic effects of VUV and AO may cause the erosion of Teflonmore severe.
文摘地面目标伪装效果评估是基于目标在探测频段下伪装前后检测概率的变化,合成孔径雷达能够全天候工作,分辨率较高,已经成为对地侦查的主要手段。对于SAR图像的目标检测概率通常采用专家判读和特征模板匹配等方式获取,这些方法主观性较强或对算法设计的依赖度较高,因此需要一种科学客观的评判方法。随着神经网络在图像识别领域的快速发展,目标特征提取和识别结果愈加准确,现采用YOLO-V3(You Only Look Once)算法模型,对MSTAR目标背景合成图像进行机器学习,自动获取目标特征进行识别。训练后的模型在测试集的识别准确度为97.87%,用训练好的模型对未伪装样本与伪装后样本分别进行检测概率计算,通过比较检测概率的变化从而量化某类目标的伪装效果。