Die Bodenkultur - Journal for Land Management, Food and Environment

D. Stajnko, M. Lakota, M. Hočevar and Z. Čmelik:

Application of thermal imaging for estimating current yield in the apple orchard

Summary

The paper describes a thermal imaging based technique for estimating number, diameter and yield of apple fruits automatically under natural condition in the orchard. A thermal camera was mounted on a tool-frame tractor equipped with a PC for acquiring apple trees images four times during the vegetation period June-September 2002. Each time twenty apple trees were recorded from both sides of the row. Video signals from the camera were digitised using frame grabber and the images were processed off-line using several image processing algorithms. Correlation coefficients (R2 from 0,78 to 0,81) were established between manually measured fruit number and estimated number based on the thermal imaging algorithm. For measuring the fruit’s diameter R2 between 0,56 and 0,65 were established and were slightly increasing according to the size development during the ripening. The R2 for current yield per tree were estimated from 0,69 to 0,79 when manual measurements were compared to thermal imaging. Keywords: thermal camera, image analysis, apple fruits, number, yield.