Poster_BirdScanMR1BirdScan radar systems are based on X-band radar and can detect even small targets such as small passerines, small bats or even insects. A BirdScan radar can easily record hundreds of thousands echos per month and this obviously calls for the automated classification into meaningful groups.
The feasibility and performance of automated classification of radar echos with Machine Learning has been demonstrated in a study from the Swiss Ornithological Institute (full-text).
At SBRS we have implemented state-of-the-art Machine Learning methods to automatically classify radar echos into meaningful groups. Our latest developments in this area were presented at the ENRAM international conference on radar aero-ecology.
Example: removal of insect echos in bird migration studies
Altitude vs. time scatter-plot showing all echos detected by BirdScan MR1 during three days. Each echo is plotted as a dot and colour-coded according to the class automatically recognized by the classification module.
Thanks to the classification module, echos from insects are excluded and only echos from birds remain and are forwarded to the MTR calculation module. Day/night periods shown as bluish/yellow bars.
At this site, insects were abundant during daytime and birds abundant during night. Only bird echos where used to compute MTR. In this example, lack of proper exclusion of insects woud have led to massive overestimation of day-time bird migration.
Are you looking for a radar solution?
Tell us about your project. We are looking forward to hearing from you.