摘要
介绍了1种探测X射线分布和剂量的方法.采用BP神经网络拟合图像灰度值与照射量率的关系曲线,实现了探测器的标定.分析了直接用CMOS图像传感器探测X射线的原理.探讨了神经网络参数选取的方法.利用LM算法优化BP神经网络,得到较为精确的拟合曲线和误差曲线,并且用测试数据验证该系统的误差性能指标.实验证明该算法能够较为精确的测量辐射的剂量信息,可应用于X射线探测器的标定.
This paper puts forward a design method of distribution and dose detection for X-ray. The application of BP neural network which fits the relationship between the gray value and the radiation exposure rate curve to calibrate the detector is proposed. The principle of direct application of CMOS in X-ray detection is analyzed. The method of neural network parame- ter selection is discussed. The LM algorithm is used to optimize the BP neural network. A more accurate fitting curve and snlaller error curve are obtained, and aset of test data is used to verify the error performance of the system. With more accu rate dose information of X-ray, this method is successfully applied to calibrate the X-ray detector, which is consistent with the expected design goals.
出处
《河南师范大学学报（自然科学版）》
CAS
北大核心
2013年第5期39-41,52,共4页
Journal of Henan Normal University(Natural Science Edition)
基金
国家自然科学基金(10905017)
河南省科技厅科技攻关重大项目(122102210434)
