We are wondering whether anybody is aware of a free/open-source system to extract and count moving signals from a videostream, with the ultimate goal to apply such a system to resolve the global conservation status of a globally threatened seabird, the Yelkouan Shearwater.
This species migrates through a narrow bottleneck (the Bosphorus, Istanbul, Turkey), where each February up to 90,000 birds have been observed. Because the migration is constant, and human observer coverage has so far been limited, an automated camera monitoring system would be useful to quantify the number of birds migrating through this sytem.
The Bosphorus is only 700 m wide at it’s narrowest point, and there are two bridges spanning the Bosphorus that are 64 m above sea level. The birds therefore fly through an archway where cameras could be placed either laterally or above (looking down from a bridge). There are not many species, and the Yelkouan shearwater is the only medium-sized, fast-flying brownish bird, so identification should be simple (for humans). The birds usually migrate in flocks (30 birds or more), where birds obscure each other, and they may fly in circles, so that birds going north need to be counted seperately from birds flying south to avoid double-counting.
The system that we would need is one that identifies and counts migrating shearwaters, which is similar to the systems developed for aerial surveying only that in our case it is not the camera platform that is moving, but the birds moving rapidly (30 km/h) through the stationary field of view of the camera.
If anybody has any suggestions or would like to collaborate on developing a system, my PhD student Dilek Sahin and me would be grateful for any help.