Acoustic bat species recognition by synergetic pattern recognition

• Recognition of bat species by their calls is difficult, for many species almost impossible (e.g. Genus Myotis)
• Considerable hearing experience with species present in the region and the acoustic response of the tools (batdetectors) applied is neccessary to be able to determine identifyable bats.
• Habitat assessment in terms of bats foraging there can only be achieved on a very exemplary base.

Our approach in search of a solution to this

• Design an automated system, which can differentiate between average search calls of all bat species (present in Switzerland, for the time being...). We use a synergetic pattern recognition algorithm for sonagram discrimination.
• Find a high speed PCMCIA data acquisition card for use in laptop computers to record ultrasound directly to memory. Working with the powerful Canary program for Macintosh computers, we develop a sound input device driver to enable Canary to record realtime with >300 kHz sampling rate.
• Design an automated system, which records and identifies signals realtime and stores classification results with time stamp for a full night.

More ? Preliminary and more detailed results are available...