利用图像相关性搜寻运动目标  

Searching Moving Objects by Image Correlation

作　　者：潘祖恒 彭青玉[1,2] 陆星 陈筱[1,2] 李丹 Pan Zuheng;Peng Qingyu;Lu Xing;Chen Xiao;Li Dan(College of Information Science and Technology,Jinan University,Guangzhou 510632,China;Sino-French Joint Laboratory for Astrometry,Dynamics and Space Science,Jinan University,Guangzhou 510632,China;Department of Physics,Jinan University,Guangzhou 510632,China)

机构地区：[1]暨南大学计算机科学系,广东广州510632 [2]暨南大学中法天体测量、动力学与空间科学联合实验室,广东广州510632 [3]暨南大学物理学系,广东广州510632

出　　处：《天文研究与技术》2023年第1期41-49,共9页Astronomical Research & Technology

基　　金：国家自然科学基金(11873026,11273014,U1431227);科技部国家重点研发计划政府间国际科技创新合作项目(2022YFE0116800);暨南大学优秀本科推免生科研创新培育计划资助.

摘　　要：在天文图像处理领域,图像减法技术应用广泛。由于不同的大气条件、曝光时间和滤光片等原因,两幅图像无法直接进行图像相减。提出一种新的图像减法算法,其本质是从统计学上利用相关性消除两幅图像之间具有相似流量分布的部分,并保留具有不同流量分布的部分。该算法执行速度快,数值稳定且局部独立。基于本文算法,开发了一套以Python语言作为接口,C语言作为底层实现的图像减法代码。使用该算法以及其他3种类似的算法处理天文图像,试验结果表明,该算法能够在短时间内找出两幅图像之间的差异,并快速定位运动目标,同时具有良好的鲁棒性和位置测量稳定性。In the field of astronomical image processing,image subtraction techniques are widely used.Due to the varying seeing conditions,different exposure times,and different filters,image subtraction cannot be performed on two images.A novel image subtraction algorithm is proposed in the paper.Its essence is based on correlation to eliminate the parts with the similar flux distribution and preserve the parts with the different flux distribution between two images in sense of statistics.The algorithm can be fast executive,is numerically stable,and is locally independent.Based on our algorithm,we have developed a set of image subtraction code with Python as the interface and C as the implementation.We use the algorithm,together with three other similar algorithms,to perform with the astronomical images.Experiments show that our algorithm can find the difference between two images and detect moving objects in a very short time while having good robustness and position measurement stability.

关 键 词：相关性 图像处理 图像减法 天体测量 运动目标 

分 类 号：P129[天文地球—天体测量]