The RaspGrade Dataset: Towards Automatic Raspberry Ripeness Grading with Deep Learning

2025-05-13

AI Summary

This study addresses the challenge of real-time, non-intrusive, five-level ripeness grading of raspberries on industrial conveyor belts. To tackle this, we propose an end-to-end visual analysis framework. Methodologically, we first construct RaspGrade—the first high-quality, instance-annotated dataset tailored for fine-grained ripeness assessment—featuring motion-captured berries, pixel-precise segmentation masks, and five-tier ripeness labels; it is publicly released on Hugging Face. Second, we design a synergistic modeling approach integrating instance segmentation with robust color- and occlusion-aware classification: leveraging Mask R-CNN for accurate fruit-level segmentation, and enhancing classification resilience against color ambiguity and partial occlusion. Experimental results validate the feasibility of five-level ripeness grading and identify color confusion and occlusion as primary bottlenecks. Our work establishes a new benchmark, introduces a novel open dataset, and proposes a principled method for intelligent agricultural sorting.

Technology Category

• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Evaluation and Analysis

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance

Abstract

This research investigates the application of computer vision for rapid, accurate, and non-invasive food quality assessment, focusing on the novel challenge of real-time raspberry grading into five distinct classes within an industrial environment as the fruits move along a conveyor belt. To address this, a dedicated dataset of raspberries, namely RaspGrade, was acquired and meticulously annotated. Instance segmentation experiments revealed that accurate fruit-level masks can be obtained; however, the classification of certain raspberry grades presents challenges due to color similarities and occlusion, while others are more readily distinguishable based on color. The acquired and annotated RaspGrade dataset is accessible on HuggingFace at: https://huggingface.co/datasets/FBK-TeV/RaspGrade.

Problem

• Develop deep learning for raspberry ripeness grading
• Address color similarities and occlusion challenges
• Create annotated dataset for industrial real-time use

Innovation

• Uses deep learning for raspberry ripeness grading
• Applies instance segmentation for fruit-level masks
• Addresses occlusion and color similarity challenges