An intelligent precision orchard pesticide spray technique based on the depth-of-field extraction algorithm

Abstract

Microsoft's Kinect system can capture the colour and depth information of a scene in real time. However, to date, there have been no known reports on the application of the Kinect system in the area of precision spray control. To obtain a relatively good spray effect, the present study integrates the advantages of colour and depth information using Microsoft's Kinect system and proposes equations for calculating the average distance between the Kinect system and a fruit tree, as well as the leaf wall area (LWA) density, to address the difficulty in estimating the dose of sprayed pesticides. First, to adjust and control the spray intensity of sprayers and the dose of sprayed pesticides, the present study proposes an equation for calculating the LWA average distance of fruit trees. The experimental results showed that the average distance was largest between the Kinect system and the bottom part or trunk of a fruit tree. A comparison with the measured distances showed that the distances calculated based on the data acquired by the Kinect system were accurate. Second, to better control the dose of sprayed pesticides, the present study proposes the concept of LWA density. Finally, the results of the experiment on peach trees, apricot trees and grapevines demonstrated that the intelligent orchard pesticide precision spray model established based on the average distance and the LWA density can improve the efficiency in spraying pesticides, reduce waste and environmental pollution, and achieve automated and precision orchard production.