Researchers from the University of California, Riverside have invented a cost-effective solution to identify insects with a 99 percent degree of accuracy that outstrips even expert efforts. The three year effort has effectively doubled the volume of available data on insects. Such an insect identification system shows great potential for addressing issues with specific insect species known to cause disease and crop damage.
Compared to larger animals, insects can be incredibly hard to identify. For starters, entomologists, or scientists that study insects, have identified over a million insect species. On top of that, some entomologists estimate that there could be yet another 30 million insect species that have yet to be discovered and/or correctly identified. Furthermore, insects can come in many shapes and sizes that may vary based upon sex, stage of development, and even environmental factors.
Traditional approaches to identifying insects involve correctly observing and interpreting many facets about the anatomy and behavior of the bug in question. Some of these clues, such as “does it have wings?” or “what kind of mouth parts does it have?” are easy to uncover. But other key identifiers may call for more skilled scrutiny such as looking for enlarged hip joints or twisting motions of the wings during flight.
In an effort to greatly simplify this laborious process, a team of computer programmers and entomologists from the University of California, Riverside collaborated on creating a new, simple, and cost-effective way to identify insects. This resulted in the invention of a laser-based sensor that was first used to identify insects based upon their flight pattern. Insects that fly through the sensor’s laser beam create a brief and partial occlusion of the laser-beam’s light. The resulting light fluctuations are then interpreted by a phototransistor as a change in current, which is subsequently amplified and filtered by a custom-made electronic board. The output of the electronics board is then used to create an MP3 file that can be downloaded to any normal computer.
Upon starting the project, the researchers began with only trying to identify six key insect species. However, as the potential of the technology became apparent, they expanded the abilities of their insect identification system to include other data besides wing-beat patterns. Upon factoring information on the time of day, the accuracy with which insects were identified jumped from 88 percent to 95 percent. And after considering data about the location of the observations, accuracy increased again to 97 percent. Researchers are currently exploring even more avenues to heighten accuracy by including information on variables such as local climate conditions and the height at which the flight pattern is observed.
The new insect identification system shows great promise for the future, but as of now entomologists need not fear for their job security quite yet. While the identification system can indeed score a whopping 99 percent on accuracy when the choice is between only two insect species, this accuracy drops precipitously the more species must be considered. With the addition of ten species, accuracy is reduced to 79 percent. Because of this, more fine-tuning will be required before this device can be expected to function as an all-in-one insect identifying solution.
By Sarah Takushi