差分CLEAN算法在编码板成像中的应用

Application of a Differential-CLEAN Algorithm in the Coded-Mask Imaging

编码板多用于硬X射线及更高能量波段的天文成像。采用交叉相关算法对编码板图像进行重建时会产生鬼像,成像质量不高。采用CLEAN算法可以有效消除鬼像,但该算法仍然无法抑制由不等间隔采样和数据缺失引入的误差。采用差分CLEAN算法对编码板图像进行重建。该算法直接在探测器数据空间进行处理,由模型数据推算出探测器上各点的数值,无需对数据进行插值等操作,能够规避上述问题,提高成像质量。选用理想点源数据加入了泊松噪声和人造坏数据点进行了模拟重建,结果显示差分CLEAN算法能够得到比CLEAN算法更低的噪声水平和更接近真实值的点源强度。IBIS是INTEGRAL卫星上的主要编码板成像设备,目前已经有大量损坏的探测器单元。用差分CLEAN算法对IBIS望远镜的数据进行了处理,与INTEGRAL标准处理软件OSA采用的CLEAN算法相比也得到了噪声水平更低的结果。

The Coded-Mask imaging method has application in astronomical imaging in hard X-ray and higher-energy bands.The cross-correlation algorithm for image reconstruction may produce ghost peaks.The CLEAN algorithm can effectively remove the ghost peaks,but still cannot avoid the error caused by the irregular data sampling or missing data points.We apply a Differential-CLEAN algorithm to reconstruct coded-mask images.This algorithm operates directly in the detector data space,and evaluates the observed data from models on each detector without data interpolation.It improves image quality by avoiding the deterioration caused by the known problems.Reconstruction results of simulated data,where ideal point-source data with the Poisson noise and artificial bad pixels are used,show that the Differential-CLEAN algorithm has better flux estimations with lower noise levels compared to the CLEAN algorithm.The IBIS,a coded-mask imager on-board the INTEGRAL satellite,has already had a considerable portion of detector units damaged.We also processed certain IBIS data with the Differential-CLEAN algorithm.Compared to the standard scientific data products produced by the INTEGRAL Offline Science Analysis（OSA） software package,which employs the CLEAN algorithm,the results of the Differential-Clean algorithm show lower noise levels.