Estimating the density of the shoot blight and rolled leaves using aerial photography from a multirotor-type drone in tea fields

Abstract

We estimated the density of shoot blight from Pestalotiopsis longiseta and rolled leaves produced by Adoxophyes honmai or Homona magnanima in tea fields using a direct counting method based on aerial photography from a multirotor-type drone. We used a DJI Phantom 4 to conduct aerial photography of shoot blight and rolled leaves. We then counted the densities of both the diseased and rolled leaves in-situ. We examined the aerial photographs on a LCD, and identified and counted the densities of both the diseased and rolled leaves. The relationship between the densities of shoot blight counted using the LCD and in-situ was R² = 0.69 and 0.76, respectively. The accuracy of estimation using the LCD was about 80%. These results suggested that estimation of the density of shoot blight based on aerial photography was effective. For the rolled leaves, the accuracy of counting using an LCD increased to about 60% from <40%, after comparing to the first count.