Harvest Forecasting
Uses weight, size and field records to support supply planning.
25-26J-016 | AI + IoT research
Export better pineapples.
With smarter grading.
PineSense integrates computer vision, sensors and IoT communication to predict harvest yield, classify ripeness, detect defects and grade pineapples for export.
Model accuracy
93%
Latency
5 ms
Defects
Low
Project overview
25-26J-016 | CSNEOne smart pipeline for export-ready pineapple decisions.
PineSense brings forecasting, machine vision and IoT control into one workflow so farmers and exporters can plan harvests, inspect fruit and separate grades with less manual guesswork.
Capture
Predict
Grade
Sort
Ripeness Detection
Classifies unripe, ripe and overripe fruit from camera images.
Defect Grading
Detects black spots, holes and fungus to assign export grades.
IoT Dashboard
Connects ESP32, sensors, MQTT and monitoring interfaces.
Product concept
From field data to automated export sorting.
Cameras, load cells and sensors capture fruit condition while AI models classify ripeness, detect defects and convert measurements into export-focused grades and planning insights.
Harvest forecasting
Ripeness detection
Export grading
Scan status
Active
Predicted grade
A+
Output
Sorted
01 - Domain
Research foundation behind PineSense.
The research focuses on reducing export losses through objective fruit assessment, real-time sensing and automated grading support.
Literature Survey
Existing work on fruit grading, YOLO-based object detection, precision agriculture and IoT monitoring guided the system design.
Research Gap
Current pineapple workflows are manual and fragmented, with no single system for harvest prediction, ripeness detection, defect grading and farm monitoring.
Research Problem
How can AI and IoT be combined to produce consistent export-quality pineapple decisions while reducing post-harvest losses?
Research Objectives
Develop an integrated system for yield forecasting, ripeness classification, defect detection, quality grading and real-time dashboard monitoring.
Methodology
Collect image, weight and sensor data, train YOLO-based models, run edge/device workflows and send results through MQTT to a monitoring dashboard.
Technologies Used
YOLO11, Python inference workflows, ESP32 firmware, MQTT, Node-RED concepts, load cells, cameras, gas sensors and dashboard interfaces support the prototype.
02 - Milestones
Research progress and implementation timeline.
Milestones track the proposal, model development, IoT integration and final research delivery for the export-quality pineapple system.
October 2025
Project Proposal
Defined the AI + IoT solution scope, objectives, datasets and export-quality grading problem.
CompletedTAF accepted
January 2026
Progress Presentation 1
Presented early YOLO models, defect classes, forecasting workflow and IoT communication progress.
CompletedPrototype 60%
March 2026
Progress Presentation 2
Showcased model results, grading logic, dashboard concepts and end-to-end integration progress.
CompletedModel results
May 2026
Final Assessment
Complete harvest forecasting, ripeness classification, defect grading and IoT dashboard workflow.
PendingFinal demo
May 2026
Viva
Demonstrate each component and explain individual research contribution.
PendingResearch defense
Project Proposal
October 2025 | Completed | TAF accepted
Progress Presentation 1
January 2026 | Completed | Prototype 60%
Progress Presentation 2
March 2026 | Completed | Model results
Final Assessment
May 2026 | Pending | Final demo
Viva
May 2026 | Pending | Research defense
03 - Documents
Project documents.
Official assessment material for the Exporting Best Quality Pineapples research project.
05 - About us
Meet the PineSense team.
Each member owns one part of the integrated AI + IoT pineapple grading workflow.
Mr. Uditha Dharmakeerthi
Lecturer, SLIIT
Mr. Ashvinda Iddamalgoda
Lecturer, SLIIT (Co-Supervisor)
HASARA MK
Harvest forecasting with weight and environmental data
GAMAGE MG
IoT integration, ESP32 control and MQTT communication
WILAWALAAARACHCHI WASH
YOLO-based pineapple ripeness classification
SAMARAKOON SMIP
Defect detection and export-quality grading logic
06 - Contact us
General project contact.
Contact the PineSense team for research discussions, prototype demonstrations or documentation related to project 25-26J-016.
pinesense.project@my.sliit.lk
Phone: +94 77 000 0000
Location: Sri Lanka Institute of Information Technology, Malabe