Online Estimation of Table-Top Grown Strawberry Mass in Field Conditions with Occlusions

2025-07-31

AI Summary

To address the challenge of non-destructive, real-time strawberry mass estimation in field-based suspended cultivation—where occlusion and variable fruit orientations degrade accuracy—this paper proposes an RGB-D vision–guided deep learning framework. First, YOLOv8-Seg enables high-precision instance segmentation. Second, an enhanced CycleGAN performs generative inpainting of occluded regions, while a tilt-correction algorithm improves robustness in projected area computation. Finally, polynomial regression maps geometric features to fruit mass. Experiments show mean absolute errors of 8.11% for isolated fruits and 10.47% under severe occlusion. The improved CycleGAN outperforms LaMa by +12.3% in pixel-area ratio and +9.6% in intersection-over-union. This work achieves contactless, online, and highly robust mass estimation under complex field conditions, providing a key technical foundation for intelligent strawberry harvesting.

Technology Category

• Computer Vision: Segmentation
• Intelligent Robots: State Estimation
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Accurate mass estimation of table-top grown strawberries under field conditions remains challenging due to frequent occlusions and pose variations. This study proposes a vision-based pipeline integrating RGB-D sensing and deep learning to enable non-destructive, real-time and online mass estimation. The method employed YOLOv8-Seg for instance segmentation, Cycle-consistent generative adversarial network (CycleGAN) for occluded region completion, and tilt-angle correction to refine frontal projection area calculations. A polynomial regression model then mapped the geometric features to mass. Experiments demonstrated mean mass estimation errors of 8.11% for isolated strawberries and 10.47% for occluded cases. CycleGAN outperformed large mask inpainting (LaMa) model in occlusion recovery, achieving superior pixel area ratios (PAR) (mean: 0.978 vs. 1.112) and higher intersection over union (IoU) scores (92.3% vs. 47.7% in the [0.9-1] range). This approach addresses critical limitations of traditional methods, offering a robust solution for automated harvesting and yield monitoring with complex occlusion patterns.

Problem

• Estimating strawberry mass under occlusion in field conditions
• Developing real-time vision-based mass estimation pipeline
• Improving accuracy for automated harvesting and yield monitoring

Innovation

• RGB-D sensing and deep learning integration
• YOLOv8-Seg for instance segmentation
• CycleGAN for occluded region completion