We have developed a set of methods to detect meteor light traces captured by all-sky CCD cameras. Operating at small automatic observatories (stations), these cameras create a network spread over a large territory. Image data coming from these stations are merged in one central node. Since a vast amount of data is collected by the stations in a single night, robotic storage and analysis are essential to processing. The proposed methodology is adapted to data from a network of automatic stations equipped with digital fish-eye cameras and includes data capturing, preparation, pre-processing, analysis, and finally recognition of objects in time sequences. In our experiments we utilized real observed data from two stations.

我们开发了一套检测全天CCD相机捕捉到的流星光迹的方法。这些摄像机在小型自动观测站（站）运行，创建了一个覆盖大片区域的网络。来自这些站的图像数据被合并到一个中央节点中。由于各站在一夜之间收集了大量数据，因此机器人存储和分析对于处理至关重要。所提出的方法适用于来自配备数字鱼眼相机的自动站网络的数据，包括数据捕获、准备、预处理、分析以及最终按时间顺序识别对象。在我们的实验中，我们利用了来自两个站点的真实观测数据。