| Titre : | Automatic date fruit sorting system based on machine learning and visual features |
| Auteurs : | Ibtissam Boumaraf, Auteur ; Abdelhamid Djeffal, Directeur de thèse |
| Type de document : | Thése doctorat |
| Editeur : | Biskra [Algérie] : Faculté des Sciences Exactes et des Sciences de la Nature et de la Vie, Université Mohamed Khider, 2024 |
| Format : | 1 vol. (128 p.) / ill., couv. ill. en coul / 30 cm |
| Langues: | Anglais |
| Mots-clés: | Convolutional Neural Networks (CNNs), Date Fruit, Image Classication, Multi-modal Data Fusion, Multi-view Imaging, Thermal Imaging, Transfer Learning, Weight Measurement. |
| Résumé : |
The Algerian date market holds signicant economic potential, ranking third in global date production. However, a substantial gap exists between its production capacity and date fruit exports. This limitation stems from slow, error-prone, and labour-intensive traditional manual sorting methods that rely on visual inspection of various quality factors. This thesis addresses these limitations by leveraging Convolutional Neural Networks (CNNs) to automate date fruit sorting. CNNs incorporate information beyond single-view visual data, creating a more ecient solution. Our novel approach utilizes a multimodal dataset that combines features from multiple fruit faces alongside thermal imaging data and weight measurements, providing a richer and more comprehensive representation of each date fruit. The thesis delves into three key contributions that explore and demonstrate the eectiveness of these CNN-based approaches. The rst contribution demonstrates the eectiveness of a multi-modal approach with CNNs, achieving 94% testing accuracy using a VGG16 model by combining all information into one visual data input. The second contribution investigates multi-modal data fusion with late fusion techniques. In Scenario I, fruits are classied based on four-view images. Scenario II extends scenario I by incorporating thermal images and weight measurements. The results highlight the signicant accuracy improvement observed when incorporating additional features in Scenario II. The nal contribution addresses the limitations of single-face analysis and small datasets. It proposes a method to combine information from multiple fruit faces and utilizes permutation functions to increase dataset size. This approach signi- cantly enhances classication accuracy, with a ne-tuned VGG16 model achieving perfect accuracy (100%) with merged four faces, highlighting the potential of data augmentation techniques to address limitations associated with limited datasets. In conclusion, this thesis demonstrates the potential of Convolutional Neural Networks (CNNs) combined with multi-modal data fusion. By leveraging information from four visual images capturing dierent faces of the date fruit, the proposed approach enhances the accuracy and richness of information about the entire fruit. This paves the way for revolutionizing automated Algerian date fruit sorting, ultimately leading to a more ecient and accurate future for the Algerian date fruit market. |
| Sommaire : |
Table of Contents vii List of Figures x List of Tables xiii 1 INTRODUCTION 1 1.1 Context . . . . . . . . . . .. . . 1 1.2 Motivation and Objectives . . . . . . . . . .. 2 1.3 Thesis Contributions . . . . . . . . . . . . . . 1.4 Thesis Structure . . . . . . . . . . . . . . . . . 4 Part I Background and Literature Review 6 2 Date Fruit Sorting 7 2.1 Introduction . . . . . . . .. . . . . 7 2.2 Overview of Date Palms . . . . . . . . . . . . . . . . . . . . 7 2.2.1 Global Date Palm Production . . . . . . . . . . . . . . 8 2.2.2 The Date Palm in Algeria . . . . . . . . . . . . . . .. 9 2.3 Comprehensive Overview of Dates . . . . . . . . . . . . . . 10 2.4 Developmental Growth Stages of Dates . . . . . . . . .. . 10 2.5 Date Fruit Classications . . . . . . . . . 11 2.6 Date Fruit Varieties in Algeria . . . . . . . . . . . . . . . . . . 12 2.6.1 Commercial Varieties of Dates . . . . . . . . . . . . 12 2.6.2 Common Dates . . . . . . . . . . . . . . . . . . . 12 2.6.3 Secondary Dates . . . . . . . . . . . . 13 2.7 Grading and Sorting Process . . . . . . . . . . . . . .. . . . 15 2.8 Quality Assessment . . . . . . . . . . . . . . . . . . . . . . 16 2.8.1 Consumer Perspective . . . . . . . . . . . . . . . .. 16 vii2.8.2 Producers Perspective . . . . . . . . . . . . 16 2.8.3 USDA Standards for Date Grades . . . . . . 16 2.8.4 UNECE Standard DDP-08 . . . . . . . . . . . . . . . 21 2.8.5 The CODEX STAN 143-1985 . . . . . . . . . . . . 22 2.9 Conclusion . . . . . . . . . . . . . . . . . . . 22 3 Machine Learning Fundamentals for Image Classication 24 3.1 Introduction . . . . . . . . . . . . . . . . . .. . . 24 3.2 Overview of Machine Learning . . . . . . . . . . . .24 3.3 Machine Learning Categories . . . . . . . . . . . . . . . . . . 26 3.3.1 Supervised Learning . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3.1.1 Supervised Machine Learning Types . . . . . . . . . . . 27 3.3.1.2 Common Supervised Learning Algorithms . . . . . . . . 27 3.3.2 Unsupervised Learning . . . . . . . . . . . . . . . . . . . . . . . . 29 3.3.2.1 Unsupervised Machine Learning Techniques . . . . . . . 29 3.3.2.2 Common Unsupervised Learning Algorithms . . . . . . . 30 3.3.3 Reinforcement Learning (RL) . . . . . . . . . . . . . . . . . . . . 31 3.3.3.1 Reinforcement Learning Types . . . . . . . . . . . . . . 31 3.3.3.2 Common Reinforcement Learning Algorithms . . . . . . 32 3.4 Deep Learning for Image Classication . . . . . . .. . . 32 3.5 Articial Neural Networks . . . . . . . . . . . . . .. . . 33 3.6 Articial Neural Network Components . . . . . . . . . . . 34 3.6.1 Hyperparameters . . . . . . . . . . . . . . . . 37 3.6.2 Optimizers . . . . . . . . . . . . . . . . . . . . . . . . 37 3.7 Types of Articial Neural Networks . . . . . . . . . . . . . 38 3.7.1 Recurrent Neural Networks (RNNs) . . . . . . . 39 3.7.2 Convolutional Neural Networks (CNNs) . . . . . . 39 3.7.3 Transfer Learning . . . . . . . . . . . . . . . . . . . . 41 3.7.3.1 Pretrained CNN Architectures . . . . . . . . 42 3.7.3.2 Fine-tuning . . . . . . . . . . . . . . .. . . . 44 3.8 Evaluation Metrics . . . . . . . . . . . . . . . . 44 3.9 Conclusion . . . . . .. . . 47 4 Literature Review 48 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . 48 4.2 Application of Articial Intelligence Systems in Agricultural Products . 48 4.2.1 Classication Using Traditional Machine Learning Techniques . . 49 4.2.2 Classication Using Deep Learning Techniques . . . . . . . . . . . 51 4.3 Date Fruit Classication: Traditional and Deep Learning Approaches . . 56 viii4.3.1 Date Fruit Classication Systems Using Machine Learning . . . . 56 4.3.2 Application of Deep Learning in Date Fruit Classication . . . . . 58 4.4 Conclusion . . . . . . . . . . . . .. . . . 63 Part II CONTRIBUTIONS 64 5 Improving date fruit sorting with a novel multimodal approach and CNN 65 5.1 Introduction . . . . . . . . . . . . . . . . . 65 5.2 The Proposed Method . . . . . . . . . . . . . . . . . . 66 5.2.1 Dataset Acquisition . . . . . . . . . . . . . . . . .. 66 5.2.2 Data Preprocessing . . . . . . . . . . . .. 68 5.2.2.1 Image Grayscale Transformation . . . . . . . 69 5.2.2.2 Image Averaging . . . . . . . . . . . . . . . . 70 5.2.2.3 Customising Image Channel Values . . . 70 5.2.3 Model Conception . . . . . . . . . . . . . . . . . . . 72 5.2.3.1 Transfer Learning Approach . . . . 72 5.2.3.2 Custom CNN Model . . . . . . 73 5.3 Results and Discussion . . . . . . . . . 74 5.4 Conclusion . . . . . . . . . . . . . . . . . . . . 83 6 Multimodal Data Fusion and Deep Learning for Automated Date Fruit Classication 84 6.1 Introduction . . . .. . . . 84 6.2 Methodology . . . . . . . . . . . . . 84 6.2.1 Data Collection . . . . . . . . . . . . . . . . . .. 86 6.2.2 Classication Scenarios . . . . . . 6.2.2.1 Scenario I: Multi-View Fusion with Deep Learning Architectures . . . . . 88 6.2.2.2 Scenario II: Multimodal Fusion with Deep Learning Architectures . . . . . . . . 89 6.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 6.3.1 Classication Results on Multi-View Data (Scenario I): . . . . . . 90 6.3.2 Classication Results on Multimodal Data (Scenario II) . . . . . . 93 6.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 7 Optimizing Date Fruit Classication Through Multi-View Imaging and Deep Learning 99 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . 99 7.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 99 ix7.2.1 Dataset Preparation . . . . . . . . . . . . . . 101 7.2.1.1 Merging Faces Step . . . . . . . . . . . . . .. 102 7.2.1.2 Permutation Step . . . . . . . . . . . . . 03 7.2.2 Training and Testing Step . . . . . . . . . . . .. . 106 7.3 Experimental result and Discussion . . . . . . . . . . . . . 109 7.4 Conclusion . . . . . . . . . . . . . . . . . . 119 8 General Conclusion 120 Bibliography 128 |
| En ligne : | http://thesis.univ-biskra.dz/id/eprint/6588 |
Disponibilité (1)
| Cote | Support | Localisation | Statut |
|---|---|---|---|
| TINF/198 | Théses de doctorat | bibliothèque sciences exactes | Consultable |
