An Efficient Approach to Beetle Damage Detection of Banana Using Deep Learning

Key points

• This study presents a lightweight framework based on YOLOv8n deep learning model for locating individual bananas and classifying their visible surface condition as healthy or unhealthy.1
• The term beetle damage in the paper title is used operationally for the locally observed scar-like surface condition; the model does not identify a specific beetle species or biological cause.1
• A custom dataset of 2,473 images containing 5,827 annotated banana instances was prepared from local-market imagery in Bangladesh.1
• The results support YOLOv8n as a compact baseline for instance-level banana surface inspection within the studied acquisition conditions.1

Sources (1)

• [1]An Efficient Approach to Beetle Damage Detection of Banana Using Deep Learning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 03:55 PM

  Visible surface damage of banana fruits by Basilepta subcostata (the banana fruit scarring beetle) can reduce their market value.

  This study presents a lightweight framework based on YOLOv8n deep learning model for locating individual bananas and classifying their visible surface condition as healthy or unhealthy.

Extractive summary: sentences quoted from the sources.