Development of real-time laser-scanning system to detect tree canopy characteristics for variable-rate pesticide application

Improving pesticide application efficiency is increasingly important in orchard spraying. In this study, a laser-scanning system was designed to acquire gridding volumes of a tree to quantify the geometry characteristics of the tree canopy in real-time. A laser-scanning sensor mounted on a linear guide was utilized to measure the structure of a target tree canopy. A computer was used to receive measurement data from the laser scanner and obtain the movement distance of the laser-scanning sensor from a controller. An algorithm written with VC# program was designed to calculate gridding volumes of trees by recognizing valid measurement data from the laser scanner. Laboratory evaluations were conducted on three kinds of regular objects, and the maximum relative errors of section volumes of the cuboid, triangular prism and cylinder objects were 3.3%, 7.9% and 9.4%, respectively, which illustrated that the algorithm could calculate the section volumes in different parts of the objects with high accuracy. A conifer tree and an apple tree were chosen to verify detecting accuracy of the laser-scanning system at variable speeds and grid sizes. The variation coefficients of total volumes for each kind of the tree were 0.078 and 0.041, respectively, which indicated that the laser-scanning system could be applied to provide the gridding volumes of different canopy densities in real-time with good reliability for guiding a variable-rate sprayer.