Uncovering implementable dormant pruning decisions from three different stakeholder perspectives

TL;DR

This study tackles the automation of dormant pruning by dissecting how horticulturists, growers, and pruners make pruning decisions. Through three field studies in Prosser, WA, involving Bing Cherries, Envy Apples, and Jazz Apples across UFO and V-trellis architectures, the authors apply grounded coding to extract a hierarchical pruning terminology and seven actionable heuristics linking global horticultural goals to specific cuts. The resulting terminology and heuristics, validated with follow-up interviews, provide a concrete framework for encoding pruning decisions into autonomous systems and highlight three horticultural contexts—environmental management, crop-load management, and replacement wood—that organize pruning decisions across cultivars. The work emphasizes the necessity of field-grounded terminology, multi-stakeholder input, and sensor-enabled quantification to enable robust robotic pruning in diverse orchard architectures, offering practical pathways for future autonomous pruning systems.

Abstract

Dormant pruning, or the removal of unproductive portions of a tree while a tree is not actively growing, is an important orchard task to help maintain yield, requiring years to build expertise. Because of long training periods and an increasing labor shortage in agricultural jobs, pruning could benefit from robotic automation. However, to program robots to prune branches, we first need to understand how pruning decisions are made, and what variables in the environment (e.g., branch size and thickness) we need to capture. Working directly with three pruning stakeholders -- horticulturists, growers, and pruners -- we find that each group of human experts approaches pruning decision-making differently. To capture this knowledge, we present three studies and two extracted pruning protocols from field work conducted in Prosser, Washington in January 2022 and 2023. We interviewed six stakeholders (two in each group) and observed pruning across three cultivars -- Bing Cherries, Envy Apples, and Jazz Apples -- and two tree architectures -- Upright Fruiting Offshoot and V-Trellis. Leveraging participant interviews and video data, this analysis uses grounded coding to extract pruning terminology, discover horticultural contexts that influence pruning decisions, and find implementable pruning heuristics for autonomous systems. The results include a validated terminology set, which we offer for use by both pruning stakeholders and roboticists, to communicate general pruning concepts and heuristics. The results also highlight seven pruning heuristics utilizing this terminology set that would be relevant for use by future autonomous robot pruning systems, and characterize three discovered horticultural contexts (i.e., environmental management, crop-load management, and replacement wood) across all three cultivars.

Figures (21)

• Figure 1: Dormant pruning has three major stakeholders: horticulturists, growers, and pruners. Each stakeholder interacts with orchards in a different way (formal studies, crop management, paid per-cut work) and possesses unique pruning knowledge acquired by formal techniques, hands-on experience, or repetition.
• Figure 2: Breakdown of our three studies based on participant ID and procedure. Participant IDs reflect the person's role, so P is a pruner, G is a grower, and H is a horticulturist. We report how long each participant participated in a specific study procedure and what cultivar and tree architecture (either Upright Fruiting Offshoot or V-Trellis) they interacted with.
• Figure 3: Example of a tagged tree with the four colored pruning tags. Each colored tag corresponds to a different pruning action. Green: Fruit spacing, Red: Self-shading or out-of-canopy growth.
• Figure 4: An example of how terminology was extracted from transcript quotes and visual cues during the in-field interviews. The discovered term --- "thinning cut" --- is represented in orange with its corresponding definition in pink. Finally, an example visual representation of the cut explained and demonstrated by the participant by pointing to the branch is shown in the left side of the image labeled "Visual Cue" and the accompanying text in purple.
• Figure 5: An example of our extracting data and open coding step for pruning heuristic analysis. We extracted a branch from a Jazz Apple video segment due to the presence of a colored tag (red). The extracted branch was added as a new in the excel sheet and visual cues filled in by one researcher in the appropriate columns for that cultivar's architecture (Apples: V-Trellis; Cherries: UFO).