摘要
在车辆辐射成像过程中,受到射线源的几何尺寸、探测器及信号放大电路响应时间、统计涨落等降质因素的影响,图像产生退化,表现为模糊与噪声增加。针对车辆辐射图像的退化问题,该文研究了辐射成像系统的退化模型,提出了利用生成对抗网络DeblurGAN的辐射图像复原方法。通过辐射图像的退化机制构造了辐射图像的特定数据集,用于训练DeblurGAN模型,利用训练好的模型去复原系统实际采集的车辆辐射图像。实验结果表明:该方法能够有效去除辐射图像的模糊与噪声,改善系统的成像质量。
In vehicle radiation imaging,the size of the gamma ray source,the response time of detector and signal amplification circuits,statistical fluctuations and other factors degrade the image with blurring and noise.A model was developed to predict the image degradation in a radiation imaging system to support a radiation image restoration method based on DeblurGAN.A set of radiation images with simulated blurring was used to train the DeblurGAN model that was then used to restore the images.The results show that this method effectively eliminates blurring and noise in radiation images to improve imaging quality.
作者
冷智颖
孙跃文
童建民
王振涛
LENG Zhiying;SUN Yuewen;TONG Jianmin;WANG Zhentao(Institute of Nuclear and New Energy Technology,Tsinghua University,Beijing 100084,China;Beijing Key Laboratory on Nuclear Detection&Measurement Technology,Beijing 100084,China)
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2022年第10期1691-1696,共6页
Journal of Tsinghua University(Science and Technology)
关键词
辐射图像
退化模型
生成对抗网络
图像复原
radiation image
degradation model
generative adversarial network
image restoration
