Estimating Light Interception of Orchard Trees Using LiDAR and Solar Models

Abstract: In farming of fruit trees it is of interest to know the light interception of the trees. Therefore, in this project, a geometric model of the trees was derived using LiDAR data and this was combined with a sky model to estimate the light interceptionof orchard trees. The light interception was estimated by first synthesising a discrete model of the hemispherical sky, which holds a measure of global lightradiation in each node. The light interception of the trees was then estimated by ray tracing from the sky, applying a radiation absorption model where rays passed the point cloud representation of the trees. Comparing the interception model to measurements of photosynthetically active radiation (PAR) underneath a tree, the qualitative agreement was high and the quantitative analysis showed a reasonable, albeit noisy, correspondence between the model output and the real world measurements. When comparing the estimations produced by the solar-geometry model and the tree volume (estimated also with LiDAR), a correspondence between the surface area of the tree and the interception was found. When comparing tree volume and light interception against actual yield numbers (total weight, average fruit weight and fruit count per tree), the observable trend was that light interception did better in predicting the average fruit size, while the volume did a better job of estimating the two others. The results were encouraging, however, because ground truth data were only available for 18 trees, future work will have to compare with a greater number of measurements over multiple growing seasons.