Autonomous lemon grading system by using machine learning and traditional image processing

Hanh, Le Duc; Bao, Danh Nguyen the

doi:10.1007/s12008-022-00926-w

Cited by 5 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most of these algorithms aim at approximate functions. In the color-based fruit classification problem, when fruits of the same shape differ only in the color threshold, the algorithm may not be as expected [32], and operation speed remains an important challenge. In order to enable the future application of an automatic cherry tomatoes quality classification system suitable for factories, we combined the recognition speed of machine learning techniques using the YOLOv4 network architecture.…”

Section: Resultsmentioning

confidence: 99%

YOLOv4-Driven Appearance Grading Filing Mechanism: Toward a High-Accuracy Tomato Grading Model through a Deep-Learning Framework

et al. 2022

View full text Add to dashboard Cite

In traditional agricultural quality control, agricultural products are screened manually and then packaged and transported. However, long-term fruit storage is challenging in tropical climates, especially in the case of cherry tomatoes. Cherry tomatoes that appear rotten must be immediately discarded while grading; otherwise, other neighboring cherry tomatoes could rot. An insufficient agricultural workforce is one of the reasons for an increasing number of rotten tomatoes. The development of smart-technology agriculture has become a primary trend. This study proposed a You Only Look Once version 4 (YOLOv4)-driven appearance grading filing mechanism to grade cherry tomatoes. Images of different cherry-tomato appearance grades and different light sources were used as training sets, and the cherry tomatoes were divided into four categories according to appearance (perfect (pedicled head), good (not pedicled head), defective, and discardable). The AI server ran the YOLOv4 deep-learning framework for deep image learning training. Each dataset group was calculated by considering 100 of the four categories as the difference, and the total numbers of images were 400, 800, 1200, 1600, and 2000. Each dataset group was split into an 80% training set, 10% verification set, and 10% test set to overcome the identification complexity of different appearances and light source intensities. The experimental results revealed that models using 400–2000 images were approximately 99.9% accurate. Thus, we propose a new mechanism for rapidly grading agricultural products.

show abstract

Section: Resultsmentioning

confidence: 99%

YOLOv4-Driven Appearance Grading Filing Mechanism: Toward a High-Accuracy Tomato Grading Model through a Deep-Learning Framework

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The research by Asriny, Rani, and Hidayatullah [16] employed a 4-layer CNN to classify oranges into five grade classes, including immature-orange, rotten-orange, and damaged-orange. Another study by Hanh and Bao [17] used a Yolov4 network to classify lemons as best, good, and bad. Darapaneni et al [18] put forward a MobileNet-based model to classify bananas as good or bad.…”

Section: Related Workmentioning

confidence: 99%

DenseNet-201 and Xception Pre-Trained Deep Learning Models for Fruit Recognition

et al. 2023

View full text Add to dashboard Cite

With the dramatic increase of the global population and with food insecurity increasing, it has become a major concern for both individuals and governments to fulfill the need for foods such as vegetables and fruits. Moreover, the desire for the consumption of healthy food, including fruit, has increased the need for applications in the field of agriculture that help to achieve better methods for fruit sorting and fruit disease prediction and classification. Automated fruit recognition is a potential solution to reduce the time and labor required to identify different fruits in situations such as retail stores during checkout, fruit processing centers during sorting, and orchards during harvest. Automating these processes reduces the need for human intervention, making them cheaper, faster, and immune to human error and biases. Past research in the field has focused mainly on the size, shape, and color features of fruits or employed convolutional neural networks (CNNs) for their classification. This study investigates the effectiveness of pre-trained deep learning models for fruit classification using two distinct datasets: Fruits-360 and the Fruit Recognition dataset. Four pre-trained models, DenseNet-201, Xception, MobileNetV3-Small, and ResNet-50, were chosen for the experiments based on their architecture and features. The results show that all models achieved almost 99% accuracy or higher with Fruits-360. With the Fruit Recognition dataset, DenseNet-201 and Xception achieved accuracies of around 98%. The good results exhibited by DenseNet-201 and Xception on both the datasets are remarkable, with DenseNet-201 attaining accuracies of 99.87% and 98.94%, and Xception attaining 99.13% and 97.73% accuracy, respectively, on Fruits-360 and the Fruit Recognition dataset.

show abstract

“…Several models and methods have been developed over the years to address these tasks. In this section, we examine the strengths, limitations, and contributions of existing models for multivariate classification and grading of fruits [7,8,9].…”

Section: Review Of Existing Models Used For Multivariate Classificati...mentioning

confidence: 99%

“…The size of the output matrix matches that of the input image sets. Next, using equations 5 and 6, the gradients (GX & GY along X & Y axis) of the smoothed image are computed to determine the intensity changes in various gradients.𝐺𝑋 = 𝐶𝑜𝑛𝑣(𝐶, 𝑆𝑜𝑏𝑒𝑙𝑋) … (5) 𝐺𝑌 = 𝐶𝑜𝑛𝑣(𝐶, 𝑆𝑜𝑏𝑒𝑙𝑌) … (6)Non-maximum suppression is used to thin out the edges after obtaining the gradient magnitudes along X and Y. Equation7is used to suppress the other non-maximum values while maintaining only the local maxima in the gradient levels.𝑆𝐺(𝑖, 𝑗) = 0 𝑖𝑓 𝐺(𝑖, 𝑗) ≤ 𝐺(𝑖 + 𝑑𝑥, 𝑗 + 𝑑𝑦) 𝑜𝑟 𝐺(𝑖, 𝑗) ≤ 𝐺(𝑖 − 𝑑𝑥, 𝑗 − 𝑑𝑦) …(7)…”

mentioning

confidence: 99%

DQLRFMG: Design of an Augmented Fusion of Deep Q Learning with Logistic Regression and Deep Forests for Multivariate Classification and Grading of Fruits

Said,

Joshi

2023

IJRITCC

View full text Add to dashboard Cite

Accurate categorization and grading of fruits are essential in numerous fields, including agriculture, food processing, and distribution. This paper addresses the need for an advanced model capable of classifying and grading fruits more effectively than existing methods. Traditional approaches are limited by their lower precision, accuracy, recall, area under the curve (AUC), and delay. In order to overcome these obstacles, the proposed model combines the capabilities of Deep Q Learning (DQL) for classification and Logistic Regression (LR) with Deep Forests for fruit grading process. Three distinct datasets were used to evaluate the model: the Kaggle - Fruits 360 Dataset, the FRont Experimental System for High throughput plant phenotyping Datasets, and ImageNet samples. In multiple respects, comparative analysis demonstrates that the proposed model outperforms existing methods. Specifically, it achieves a remarkable 4.9% improvement in precision, 5.5% improvement in accuracy, 4.5% improvement in recall, 3.9% improvement in AUC, and an 8.5% reduction in delay levels. Utilizing the strengths of both DQL and LR with Deep Forests, the proposed model achieves its superior performance. DQL, a technique for reinforcement learning, provides the ability to learn and make decisions based on the feedback from the environment. By combining DQL and LR, the classification accuracy is improved, allowing for the precise identification of fruit varieties including Mango, Apple. Papaya, etc. In addition, Deep Forests, a novel framework for ensemble learning, is utilized for fruit grading. Deep Forests utilizes decision trees to effectively capture complex patterns in the data, allowing for dependable and robust fruit grading. Experimental findings indicate that the combination of DQL and LR with Deep Forests yields remarkable performance improvements in fruit classification and grading tasks. Improved precision, accuracy, recall, AUC, and delay indicate the model's superiority over existing methods. This research contributes to the field of fruit classification and grading by developing a sophisticated model that can support a variety of applications in the agriculture, food processing, and distribution industries.

show abstract

Autonomous lemon grading system by using machine learning and traditional image processing

Cited by 5 publications

References 20 publications

YOLOv4-Driven Appearance Grading Filing Mechanism: Toward a High-Accuracy Tomato Grading Model through a Deep-Learning Framework

YOLOv4-Driven Appearance Grading Filing Mechanism: Toward a High-Accuracy Tomato Grading Model through a Deep-Learning Framework

DenseNet-201 and Xception Pre-Trained Deep Learning Models for Fruit Recognition

DQLRFMG: Design of an Augmented Fusion of Deep Q Learning with Logistic Regression and Deep Forests for Multivariate Classification and Grading of Fruits

Contact Info

Product

Resources

About