U.S. flag

An official website of the United States government

Dot gov

Official websites use .gov
A .gov website belongs to an official government organization in the United States.


Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.


Main content area

Use of unmanned aircraft systems (UAS) and multispectral imagery for quantifying agricultural areas damaged by wild pigs

Justin W. Fischer, Kelsey Greiner, Mark W. Lutman, Bryson L. Webber, Kurt C. Vercauteren
Crop protection 2019 v.125 pp. 104865
Sus scrofa, agricultural land, corn, crop damage, crops, income, learning, monitoring, multispectral imagery, remote sensing, swine, unmanned aerial vehicles, Missouri
Wild pigs (Sus scrofa) cause extensive damage to agricultural crops, resulting in lost production and income. A major challenge associated with assessing damage to crops is locating and quantifying damaged areas within agricultural fields. We evaluated a novel method using multispectral high-resolution aerial imagery, collected from sensors mounted on unmanned aircraft systems (UAS), and feature extraction techniques to detect and map areas of corn fields damaged by wild pigs in southern Missouri, USA. Damaged areas were extracted from ortho-mosaics using visible and near-infrared band combinations, an object-based classification approach, and hierarchical learning cycles. To validate estimates we also collected ground reference data immediately following flights. Overall accuracy of damage estimates to corn fields were similar among band combinations evaluated, ranging from 74% to 98% when using visible and near-infrared information, compared to 72%–94% with visible information alone. By including near-infrared with visible information, though, we found higher average kappa values (0.76) than with visible information (0.60) alone. We demonstrated that UAS are an appropriate platform for collecting high-resolution multispectral imagery of corn fields and that object-oriented classifiers can be effectively used to delineate areas damaged by wild pigs. The proposed approach outlines a new monitoring technique that can efficiently estimate damage to entire corn fields caused by wild pigs and also has potential to be applied to other crop types.