Abstract: After the investigation of the characteristics of celestial background image sequence，an image sequence compression method is proposed.First，estimate the background of the celestial background image sequence and filter the background noise with region self-adaptive filter.Then use the run length encoding method to encode the filtered background.Finally use morphology method to extract the moving target，calculate the target’s position in each frame and add this information to the codes of background.Experiments show that this method has great capability of compressing the celestial background image sequence.The recovered images maintain the basic information of the original raw images and have good impression which is convenient for application processing.

Key words: celestial background, background estimation, region self-adaptive filtering, run length encoding, code synthesizing

摘要： 在分析星空背景图像的特点基础上，提出了一种适用于星空背景图像序列的压缩编码方法。算法首先估计出图像序列的星空背景，采用基于自适应滤波的区域门限分割算法，在保持背景图像中星体灰度的基础上去除噪声，再对分割后的背景图像进行游程编码；其次，采用数学形态学的方法提取出运动目标图像，并计算出运动目标图像在图像序列每一帧中的位置，最后将运动目标图像和它在每帧中的位置编码嵌入到背景图像编码流，形成最终的星空背景图像序列编码结果。实验表明，该方法对于星空背景图像序列具有很强的压缩能力，恢复出的图像序列消除了噪声影响并保留了原始图像的基本信息，同时具有较好的视觉效果，便于后续的应用处理。

关键词: 星空背景, 背景估计, 区域自适应滤波, 游程编码, 码流合成