The ROAD to discovery: Machine-learning-driven anomaly detection in radio astronomy spectrograms

Home

All issues

Volume 680 (December 2023)

A&A, 680 (2023) A74

Full HTML

Open Access

Table 1

Categorisation of data processing, electronic, astronomical, and environmental anomalies in the ROAD dataset.

Category	Description	Band	Polarisation	Occurrence rate	# Samples
Normal	All non-characterised effects	Both	All	–	4687
Data processing
First order data loss	Data loss from consecutive time and/or frequency channels	Both	All	0.02	146
Second order data loss	Data loss from single frequency and/or single time channels	Both	All	0.04	283
Electronic systems
High noise element	High power disturbances caused by miscellaneous events	Both	All	0.01	88
Oscillating tile	Amplifier going into oscillation	High	All	0.01	56
Astronomical events
Source in side-lobes	A-team source passing through side-lobes	High	All	0.06	446
Galactic plane	Galactic plane passing through the main lobe of the antenna	Both	Cross	0.08	550
Solar storm	Strong emissions from the sun	Low	All	0.02	147
Environmental effects
Lightning	Lightning storm	Both	All	0.06	389
Ionospheric RFI reflections	RFI reflected from the ionosphere	Low	All	0.04	261

Notes. A-team sources refer to the four brightest persistent radio sources in the northern sky. We note that each spectrogram may contain multiple anomalies; hence, the number of samples stated is greater than the overall dataset size.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.