CCD图像拼接试验

An Experiment of CCD Frame-Combination

为了获得大视场的高精度天文图像,实施了一种望远镜CCD图像的拼接方法。从原始图像到最终合成图像的坐标转换采用了结合星表UCAC4的六常数模型。不同于硬件级的拼接,使用逐个像素的灰度值再分配的方案进行图像融合。进一步采用云南天文台2.4 m望远镜拍摄的CCD图像进行了试验。结果表明,该算法可以产生较高质量的大视场CCD图像,可以直观地发现运动目标,暗星信噪比有显著改善。高精度的图像拼接还与原始数据扭曲改正的预处理密不可分。与未做扭曲改正相比,图像拼接的位置精度提高了约一倍(约0.02 pixel)。

In order to obtain astronomic images with a wide field of view and high precision,a method of combining a few CCD（ Charge Coupled Device） frames taken by a telescope is performed in this thesis. Peng et al.＇ s geometric distortion model is used to correct original images during the pre-processing phase.Coordinate transformation from original images to the final combined frame uses a six-parameter model together with the Catalog UCAC4. Unlike hardware image mosaic,the gray value of each pixel in the original frame needs to be redistributed to the pixels of a unique reference frame. The algorithms of this frame-combination are described in detail in this article. Furthermore,this method has been tested with many CCD frames taken by the 2. 4m Telescope at Yunnan Observatories. Results show that this method can yield a quite good CCD frame with a wide field of view. Specifically,moving objects can be detected easily; in terms of faint ones,there is significant improvement in their SNR（ Signal to Noise Ratio） as well. Test results also show that the star positional errors introduced by the frame-combination are reduced to about 0. 02 pixels when raw CCD frames are corrected for their geometry distortions. This experiment therefore confirms the validity of Peng et al.＇s geometric distortion solution and the necessity of correcting geometric distortions.