AI Pruning!
AI-Supported Fruit Tree Pruning
Plant pruning is a demanding, critical, and expensive task in fruit growing that is often performed by experts or the fruit growers themselves, as mistakes affect the perennial crop in the long term. However, since skilled workers and time are lacking, 'AI Pruning' aims to help overcome these bottlenecks.
With 'AI Pruning', untrained workers can perform pruning professionally and quickly. A smartphone is used - which can be extended with smart glasses - that analyzes the plants and displays precise instructions for pruning directly. This can close a significant innovation gap between technology and practice.
Current News
The 'AI Pruning' project’s web app makes it easier for our developers to understand the model generation process and identify issues. It illustrates the various steps involved in creating 3D models. First, it compares the camera positions in the generated 3D model with the recorded keyframes selected from the video footage.
Next, the generation of the graph is shown and is checked for accuracy against the generated point cloud. Finally, the cutting positions and the corresponding rules are displayed in a table alongside the graph. In addition, the generated cutting graph can be compared here with the original graph.
Project Timeline
02.2026 Achieved stable tracking in the lab
12.2025 Start of the 2025/2026 recording season
10.2025 Cutting suggestions are displayed in the model
09.2025 Basis for precise cutting instructions created!
06.2025 Networking in Mestre at the EU-CAP seminar
05.2025 Precise 3D reconstruction of trees to be pruned
04.2025 Successful semantic segmentation
03.2025 AI Pruning presents itself to the industry
01.2025 Mass recordings in South Tyrolean production facilities
11.2024 Winter pruning, image recording with recording app of orchard meadows
03–10.2024 Development of the recording app
02.2024 Start of image material annotation
01.2024 Project start: early start of measures; winter pruning and image recording