Banana Ripeness Detection and Smart Environmental Monitoring System Using YOLOv8, Random Forest, and IOT Sensors

This paper presents the design and implementation of a Banana Ripeness Detection and Smart Environmental Monitoring System that integrates deep learning, machine learning, and Internet of Things (IoT) technologies to provide an automated, non-destructive, and real-time solution for fruit quality assessment and environmental condition management. The proposed system comprises two integrated modules. The first module employs a YOLOv8 deep learning model trained on a Roboflow-prepared dataset to classify banana images into five categories: Unripe, Semi-ripe, Ripe, Overripe, and Disease Detected. Upon detection of a critical ripeness stage or disease, the system automatically dispatches an email notification via SMTP. The second module performs continuous environmental monitoring using four IoT sensors (MQ135, DHT11, soil moisture, and pH sensor), analysed by a Random Forest classifier to determine Normal or Abnormal conditions. In abnormal cases, SMS alerts are sent via a GSM module and an irrigation pump is activated automatically. The YOLOv8 model achieved a mean Average Precision ([email protected]) of 89.6% and the Random Forest classifier achieved 96.8% overall accuracy. The system provides a scalable, cost-effective, and practical framework for precision agriculture.