Data from: Acoustic detection and classification of three stored product insect pests
Insect pests destroy large quantities of stored products, such as raw grain, flour, and retail products, resulting in significant losses. Traditional monitoring for these pests results in a lag between first infestation and detection. However, advances in sensor technology and computing have allowed for early-warning passive monitoring and automated detection. Acoustic detection of grain pests shows great promise, but accurate classification of different species at different levels of infestation is still needed. In the present study, the substrate-borne vibrations (e.g., sounds produced by arthropod activity on the commodities) of three insect species, Rhyzopertha dominica, Tribolium castaneum, and Sitophilus zeamais were recorded at three densities (5, 20, and 40 adults/kg grain) in 200 g of whole wheat. We compared two deployment mechanisms for the acoustic sensor: attached to a metal plate on top of the grain or a long screw inserted into the grain bulk. We found deployment on the metal plate was more sensitive. Additional tests with the metal plate indicated that the number of vibrations recorded and their frequency differed across both insect density and species. The duration of vibrations and the interval between them also varied for each species, with the highest number and shortest interval produced by T. castaneum. These characteristics were used to build linear discriminant models to classify species from unlabeled recordings. The best performing models predicted insect species with an accuracy of 80%. These results add to a growing body of literature illustrating the utility of passive acoustic monitoring for insects in stored products.The Beetle_Vibration_Data_Key.txt corresponds to the Beetle Vibration Data.csv to describe the column headers. Full description of data processing can be found in the accompanying manuscript.
Resources
2 resources available
-
Beetle Vibration Data.csv
CSV -
Beetle_Vibration_Data_Key.txt
TXT
Find Related Datasets
Search by Tags
Click any tag below to search for similar datasets
Complete Metadata
| @type | dcat:Dataset |
|---|---|
| accessLevel | public |
| accrualPeriodicity | irregular |
| bureauCode |
[ "005:18" ] |
| contactPoint |
{ "fn": "Gerken, Alison R.", "hasEmail": "mailto:alison.gerken@usda.gov" } |
| description | <p dir="ltr">Insect pests destroy large quantities of stored products, such as raw grain, flour, and retail products, resulting in significant losses. Traditional monitoring for these pests results in a lag between first infestation and detection. However, advances in sensor technology and computing have allowed for early-warning passive monitoring and automated detection. Acoustic detection of grain pests shows great promise, but accurate classification of different species at different levels of infestation is still needed. In the present study, the substrate-borne vibrations (e.g., sounds produced by arthropod activity on the commodities) of three insect species, <i>Rhyzopertha dominica</i>, <i>Tribolium castaneum</i>, and <i>Sitophilus zeamais </i>were recorded at three densities (5, 20, and 40 adults/kg grain) in 200 g of whole wheat. We compared two deployment mechanisms for the acoustic sensor: attached to a metal plate on top of the grain or a long screw inserted into the grain bulk. We found deployment on the metal plate was more sensitive. Additional tests with the metal plate indicated that the number of vibrations recorded and their frequency differed across both insect density and species. The duration of vibrations and the interval between them also varied for each species, with the highest number and shortest interval produced by <i>T. castaneum</i>. These characteristics were used to build linear discriminant models to classify species from unlabeled recordings. The best performing models predicted insect species with an accuracy of 80%. These results add to a growing body of literature illustrating the utility of passive acoustic monitoring for insects in stored products.</p><p dir="ltr">The Beetle_Vibration_Data_Key.txt corresponds to the Beetle Vibration Data.csv to describe the column headers. Full description of data processing can be found in the accompanying manuscript. </p> |
| distribution |
[ { "@type": "dcat:Distribution", "title": "Beetle Vibration Data.csv", "format": "csv", "mediaType": "text/csv", "downloadURL": "https://ndownloader.figshare.com/files/65647578" }, { "@type": "dcat:Distribution", "title": "Beetle_Vibration_Data_Key.txt", "format": "txt", "mediaType": "text/plain", "downloadURL": "https://ndownloader.figshare.com/files/65647593" } ] |
| identifier | 10.15482/USDA.ADC/32663844.v1 |
| keyword |
[ "BEHAVIORAL ECOLOGY", "acoustic monitoring", "bioacoustics", "bulk storage", "integrated pest management", "stored product", "vibration" ] |
| license | https://creativecommons.org/publicdomain/zero/1.0/ |
| modified | 2026-06-26 |
| programCode |
[ "005:040" ] |
| publisher |
{ "name": "Agricultural Research Service", "@type": "org:Organization" } |
| temporal | 2024-10-15/2025-09-01 |
| title | Data from: Acoustic detection and classification of three stored product insect pests |
