nowcast developed a new “noise reduction” algorithm that uses state-of-the-art learning algorithms to reliably identify EM noise.
In consequence of technological development of human society, the number of technical devices which transmit electro-magnetic (EM) signals in all frequency bands increasingly spreads over populated areas. This represents a potential issue for all procedures which measure natural EM signals since these are superimposed by these artificial EM signals.
Taking into account this fact, and also to improve the already high quality of LINET data, nowcast developed an innovative procedure to minimize the effect of man-made EM noise on LINET lightning detection.
This new “noise reduction” algorithm uses state-of-the-art learning algorithms to reliably identify EM noise. This enables its removal from the lightning information. The algorithms recognize and reject these noise signals from the measurement at a very early stage of the process chain, more precisely: already at the LINET sensor.
Noise reduction decreases the data volume which has to be transmitted. This leads to a reduced data volume which has to be processed and evaluated. The advantages are obvious: It saves time so that the performance of real time processing is improved. Furthermore, it reduces load on the technical infrastructure and cost of maintenance. But most important of all: It minimizes potential influence of EM noise.