Comparison thresholding method for migratory raptors counting system accuracy

摘要

The spread of avian influenza has caused severe deaths in poultry and humans. It is important to understand migratory raptors flyways by monitoring their migration patterns. Ground counting is the oldest and easiest method to count the population of migratory raptors. The observers recognize the species of migratory raptors based on size, colour and song and flying characteristic. However, ground counting of big flocks of migratory raptors often faces difficulties due to the raptors' mobility and size. When the size of the flocks increases, ground counters will tend to produce more errors in their counting. Besides, ground counters will feel tired and easily lost their concentration when waiting for the migratory raptors for a long period of time. Therefore, the accuracy of this ground count method is questionable. This paper proposed the usage of optical remote sensing coupled with the image processing techniques to study the migratory raptors pathways. During the development of the software system, different image processing techniques were evaluated and enhanced for developing this monitoring system. The proposed mathematical thresholding technique is able to segment the raptors image from the background effectively.