Collaborative real-time vision-based device for olive oil production monitoring

TL;DR

This work tackles automatic visual screening of olives to prevent rocks and debris from entering an olive grinder, reducing equipment damage and ensuring oil quality. It presents a low-cost, real-time system that uses a downward-facing camera, an embedded YOLOv8-based detector with a small-object emphasis, and a pan-tilt laser guidance head to alert and localize foreign objects. The authors built real-world datasets from an operational grinder, applied tiling and data augmentation (640×640 patches) with SAHI, achieving a best precision of 85.3% and recall of 52.4% on unseen data, highlighting both the promise and the need for further data collection. The approach demonstrates practical impact by enabling automatic monitoring and rapid intervention in entry-level olive-processing setups, potentially reducing downtime and preserving EVOO quality.

Abstract

This paper proposes an innovative approach to improving quality control of olive oil manufacturing and preventing damage to the machinery caused by foreign objects. We developed a computer-vision-based system that monitors the input of an olive grinder and promptly alerts operators if a foreign object is detected, indicating it by using guided lasers, audio, and visual cues.

Figures (10)

• Figure 1: A washing unit used to automatize the preprocessing of the olives
• Figure 2: Concept of the proposed solution
• Figure 3: Setup used to acquire the dataset
• Figure 4: Example dataset image. Three rocks can be seen here
• Figure 5: Tiling a source image into six smaller patches