Articles
Analysis of the leaf-area-to-fruit ratio in apple tree training systems using a mobile LiDAR laser scanner
Article number
1457_58
Pages
465 – 472
Language
English
Abstract
In recent decades, the slender spindle has become the standard tree-training system in commercial apple orchards.
Yet, alternative systems, such as two-dimensional fruiting walls, may enable more profitable production and better align with emerging mechanization opportunities.
Fruit quality strongly depends on the vegetative growth of the tree.
Since trees vary in leaf area and associated fruit-bearing capacity, the number of fruits per tree alone is insufficient for predicting performance potentials.
Therefore, to explore the yield potential, we investigated the leaf-area-to-fruit ratio just before harvest by means of light detection and ranging (LiDAR) in an experimental apple orchard near Bonn, Germany.
The experimental orchard consisted of trees of Malus × domestica Borkh. ‘Elstar’ (n=30), which were trained as slender spindle (n=10), upright biaxis trees (n=10), and Drapeau trees with three axes (n=10). A mobile LiDAR laser scanner (R2000, Pepperl+Fuchs, Mannheim, Germany) was mounted on a tractor and driven parallel to the tree rows at constant speed (0.5 km h‑1) to collect three-dimensional data along the row.
The geometric features of linearity, curvature, and the LiDAR’s apparent reflectance intensity were derived from the 3D point cloud and used to distinguish leaves from other plant parts, allowing estimation of the leaf area of each tree (LALiDAR). The LALiDAR measurements were performed 114 days after full bloom (DAFB), when the leaf area was also manually measured (LAManual) via defoliation.
In parallel, fruit diameter, color, and crop load per tree were measured during harvest at 120 DAFB (first picking) and 130 DAFB (second picking), allowing estimation of the leaf-to-fruit ratio for each tree.
Yet, alternative systems, such as two-dimensional fruiting walls, may enable more profitable production and better align with emerging mechanization opportunities.
Fruit quality strongly depends on the vegetative growth of the tree.
Since trees vary in leaf area and associated fruit-bearing capacity, the number of fruits per tree alone is insufficient for predicting performance potentials.
Therefore, to explore the yield potential, we investigated the leaf-area-to-fruit ratio just before harvest by means of light detection and ranging (LiDAR) in an experimental apple orchard near Bonn, Germany.
The experimental orchard consisted of trees of Malus × domestica Borkh. ‘Elstar’ (n=30), which were trained as slender spindle (n=10), upright biaxis trees (n=10), and Drapeau trees with three axes (n=10). A mobile LiDAR laser scanner (R2000, Pepperl+Fuchs, Mannheim, Germany) was mounted on a tractor and driven parallel to the tree rows at constant speed (0.5 km h‑1) to collect three-dimensional data along the row.
The geometric features of linearity, curvature, and the LiDAR’s apparent reflectance intensity were derived from the 3D point cloud and used to distinguish leaves from other plant parts, allowing estimation of the leaf area of each tree (LALiDAR). The LALiDAR measurements were performed 114 days after full bloom (DAFB), when the leaf area was also manually measured (LAManual) via defoliation.
In parallel, fruit diameter, color, and crop load per tree were measured during harvest at 120 DAFB (first picking) and 130 DAFB (second picking), allowing estimation of the leaf-to-fruit ratio for each tree.
Publication
Authors
L. Zimmermann, N. Tsoulias, E. Luedeling
Keywords
apple, training system, LiDAR, leaf-area-to-fruit ratio, precision horticulture
Online Articles (96)