Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Convolutional Neural Networks for Palm Fruit Maturity Classification

Mingqiang Han, Chunlin Yi

TL;DR

The paper tackles automated palm fruit maturity assessment to optimize harvest timing and oil yield. It evaluates a shallow CNN baseline and transfer learning with ResNet50 and InceptionV3 on a public, variation-rich palm fruit dataset to classify five maturity levels, using an 80/20 train/test split. Results show ResNet50 achieving the strongest test performance (~86%), with InceptionV3 close behind, while the baseline performs substantially worse; EfficientNet underperforms in generalization. The work demonstrates the viability of end-to-end deep learning for real-time harvest decision support without extensive image preprocessing, and suggests future work on cross-variety generalization and alternative architectures.

Abstract

To maximize palm oil yield and quality, it is essential to harvest palm fruit at the optimal maturity stage. This project aims to develop an automated computer vision system capable of accurately classifying palm fruit images into five ripeness levels. We employ deep Convolutional Neural Networks (CNNs) to classify palm fruit images based on their maturity stage. A shallow CNN serves as the baseline model, while transfer learning and fine-tuning are applied to pre-trained ResNet50 and InceptionV3 architectures. The study utilizes a publicly available dataset of over 8,000 images with significant variations, which is split into 80\% for training and 20\% for testing. The proposed deep CNN models achieve test accuracies exceeding 85\% in classifying palm fruit maturity stages. This research highlights the potential of deep learning for automating palm fruit ripeness assessment, which can contribute to optimizing harvesting decisions and improving palm oil production efficiency.

Deep Convolutional Neural Networks for Palm Fruit Maturity Classification

TL;DR

The paper tackles automated palm fruit maturity assessment to optimize harvest timing and oil yield. It evaluates a shallow CNN baseline and transfer learning with ResNet50 and InceptionV3 on a public, variation-rich palm fruit dataset to classify five maturity levels, using an 80/20 train/test split. Results show ResNet50 achieving the strongest test performance (~86%), with InceptionV3 close behind, while the baseline performs substantially worse; EfficientNet underperforms in generalization. The work demonstrates the viability of end-to-end deep learning for real-time harvest decision support without extensive image preprocessing, and suggests future work on cross-variety generalization and alternative architectures.

Abstract

To maximize palm oil yield and quality, it is essential to harvest palm fruit at the optimal maturity stage. This project aims to develop an automated computer vision system capable of accurately classifying palm fruit images into five ripeness levels. We employ deep Convolutional Neural Networks (CNNs) to classify palm fruit images based on their maturity stage. A shallow CNN serves as the baseline model, while transfer learning and fine-tuning are applied to pre-trained ResNet50 and InceptionV3 architectures. The study utilizes a publicly available dataset of over 8,000 images with significant variations, which is split into 80\% for training and 20\% for testing. The proposed deep CNN models achieve test accuracies exceeding 85\% in classifying palm fruit maturity stages. This research highlights the potential of deep learning for automating palm fruit ripeness assessment, which can contribute to optimizing harvesting decisions and improving palm oil production efficiency.

Paper Structure

This paper contains 17 sections, 2 equations, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methods for Classifying Palm Fruit Ripeness
  4. Dataset
  5. Our Methodology
  6. Proposed Method
  7. CNN Deep Learning
  8. ResNet50
  9. InceptionV3
  10. Experiments
  11. Evaluation Metrics
  12. Baseline Model
  13. ResNet50 Model
  14. InceptionV3 Model
  15. Comparisons of Models
  16. ...and 2 more sections

Figures (7)

  • Figure 1: Sample images from the dataset are grouped into five ripeness classes: unripe, under-ripe, medium-ripe, ripe, and overripe. These images exhibit significant variations in scale, angle, and illumination, with some palm bunches partially covered by green bags.
  • Figure 2: The ResNet architecture
  • Figure 3: The InceptionV3 architecture and the Inception module
  • Figure 4: Summary of the baseline model architecture.
  • Figure 5: Classification report and confusion matrix, where classes 0, 1, 2, 3, and 4 correspond to medium-ripe, over-ripe, ripe, under-ripe, and unripe stages, respectively. Figures (a), (b), and (c) represent the results for the baseline model, ResNet50, and InceptionV3, respectively.
  • ...and 2 more figures