Evaluating the ability to detect foreign objects in crops using laser range scanners mounted on agricultural vehicles

Abstract

The general objective of this work was to evaluate the effectiveness of a Laser Measurement Systems (LMS) mounted on an agricultural vehicle at detecting foreign objects in standing crops such as Hay, Wheat, Soy, and Oats. More specifically, to evaluate the effectiveness of various algorithms and evaluating the affect of various test parameters. A SICK LMS 291-S14 scanner was placed on an agricultural tractor to scan different standing crops in which standard test objects were placed. A high rate of detection was found for objects that were significantly taller than crops. Crop density or foliage cover had a negative impact on the detection rate for shorter test objects. Increasing vehicle speed was also found to reduce detection rates due to lower field scan resolution. The average height and density methods had greater success rates of 72.4% and 49.4%. The discontinuity and connectivity methods had a success rate of 20.6% and 18% respectively. This system, with the conjunction of other safety systems, may be useful for ensuring safe field operation of autonomous agricultural vehicles.