🌟 Project Overview: Instigates an AI-based Yield Phenotyping Framework for fruit classification and grading.
🍎 Fruit Characteristics: Focuses on homogeneity in color, shape, and size. Addresses challenges in grading and segmentation due to these uniform qualities.
🎯 Primary Objective: Establishes a real-time, non-destructive system for automating fruit segmentation, grading, and classification.
📊 Datasets and Analysis: Utilizes two different datasets. Conducts a detailed analysis of deep learning models: AlexNet and ResNet.
🍇 Fruit Diversity: Incorporates various fruit types into the prototype. Records videos of each fruit and sends them to a local server.
🤖 Deep Learning Models: Predict fruit grades. Conveyor belt system sorts fruits into baskets based on their grade.
🖼️ Image Preprocessing: Applies techniques to precisely isolate fruits in each image.
📈 Evaluation Metrics: Considers accuracy, precision, recall, and f1-score to evaluate models.
🚀 Model Performance: ResNet architecture achieves peak accuracy, precision, and recall of 99%.
📑 Research Presentation: Details the prototype’s architecture, methodologies, and experimental results. Provides an automated solution to an existing challenge in Agri-field management.