© Marcel Burkhardt
Radar surveillance has been used to quantify bird movements across the airspace since more than 50 years. For a proper quantification it is essential that the surveyed volume is known. Unfortunately, the surveyed volume (= detection range) depends on the size of a target and thus, measuring the target size is essential. In addition, there is a need to differentiate echoes from different animals (birds – bats – insects). Existing algorithms (random forest classifier) are able to classify radar signals from different targets automatically into categories. However, there are still obvious patterns in the signal not yet used, which should allow a refinement of the categories.
The aim of this study is to improve the parameterisation of the intermittent wing-flapping patterns of passerine to allow using flapping and non-flapping phases for automatic classification. In addition or as a separate study, echo signature patterns from bats could be analysed to determine potential parameters for differentiating bats and birds.
There are more than a million echo signatures available, automatically classified into passerine-type, wader-type, unknown birds, insects and others. The student must have a good mathematical/ statistical background, good experience in R-programming and a high motivation to dive into a complex signal processing problem.
Swiss Ornithological Institute
Dr. Felix Liechti
Tel: +41 41 462 97 82