Automating aquatic insect identification through pattern recognition /
Aquatic insect population counts can be a good indicator of the health, or water quality, of rivers and streams. Stoneflies (Plecoptera) are particularly susceptible to pollution in streams. However, today?s current method for obtaining these population counts requires biologists to examine individual specimens under a microscope for identification to the species level. This manual method has proven to be a cumbersome and inefficient way to monitor stream and river health. In response, an aquatic insect imaging device was designed, constructed and tested in an attempt to speed and automate the process of insect identification. The device was specifically designed to handle various species of stoneflies immersed in a fluid medium. Orientation methods are incorporated to acquire appropriate digital images useful for pattern recognition. This project suggests an alternative method for aquatic insect identification that could lead to more efficient biomonitoring.
School:Oregon State University
School Location:USA - Oregon
Source Type:Master's Thesis
Keywords:image analysis pattern perception aquatic insects biological monitoring
Date of Publication: