Utilizing Fine-Tuned YOLOv8 Deep Learning Model for Greenhouse Capsicum Detection and Growth Stage Determination
Ayan Paul, Ambuj, Harsh Nagar, Rajendra Machavaram
Abstract
Efficient monitoring and management of greenhouse crops are essential for optimizing agricultural practices. This research investigates the application of a fine-tuned YOLOv8 deep learning model for the simultaneous tasks of detecting capsicum within a greenhouse environment and determining their growth stages. To enable accurate capsicum detection and growth stage determination, the model is trained on a diverse dataset comprising various developmental phases of capsicum plants. Image data were collected from a greenhouse using a smartphone equipped with an AI camera. Among the different YOLOv8 model variants, YOLOv8m and YOLOv8n emerged as superior choices for capsicum detection and growth stage determination, achieving mean average precision (mAP) scores of 0.952 and 0.751, respectively. These model variants exhibited exceptional efficiency in terms of detection speed and accuracy. This innovative approach has the potential to significantly enhance crop yield, optimize resource utilization, and improve overall greenhouse productivity.