Poster details

Recognition of bad pixels on charge-coupled devices through statistical analysis of event lists

Presenter: Michael Wille

We present an algorithm for the detection of defective pixels on charge-coupled devices (CCDs) by statistical means. For this purpose we show a way to organize and divide the event data into logical chunks which the recognition algorithm can operate on afterwards. We demonstrate the results of our tests with real event data taken by XMM-Newton and discuss the performance of the detection tools which prove to securely detect bad pixels with a mean signal of about 2*sigma above background level. As the reliable detection of bad pixels is of great importance for instrument health checks, we conclude that our toolset should not be used in a completely autonomous environment, but under regular human supervision.

Jörn Wilms
Ingo Kreykenbohm
Christoph Großberger
Christian Schmid
Johannes Hölzl

Back to listing
We disclaim any and all responsibility or liability for the content of this website. Use latest version of your browser to avoid rendering issues.