Automated detection of meteors

Datum konání: 27.03.2015
Přednášející: Tomáš Suk
Odpovědná osoba: Kotera

I propose a new detection technique based on statistical characteristics of images in the video sequence.These characteristics displayed in time enable to catch any bright track during the whole sequence. The method is applied to the image datacubes that are created from camera pictures of the night sky.
Meteor flying through the Earth's atmosphere leaves a light trail lasting a few seconds on the sky background. A special technique is developed to recognize this event automatically in the complete observed video sequence.For further analysis leading to the precise recognition of object I suggest to apply Fourier and Hough transformations.