Automating Tomato Ripeness Classification and Counting with YOLOv9

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(4)

Published: Jan. 1, 2024

This article proposes a novel solution to the long-standing issue of ripe (or manual) tomato monitoring and counting, often relying on visual inspection, which is both time-consuming, requires lot labor prone inaccuracies. By leveraging power artificial intelligence (AI) image analysis techniques, more efficient precise method for automating this process introduced. approach promises significantly reduce requirements while enhancing accuracy, thus improving overall quality productivity. In study, we explore application latest version YOLO (You Only Look Once), specifically YOLOv9, in classification ripeness levels counting tomatoes. To assess performance proposed model, study employs standard evaluation metrics including Precision, Recall, mAP50. These provide valuable insights into model's ability accurately detect count tomatoes real-world scenarios. The results indicate that YOLOv9-based model achieves superior performance, as evidenced by following metrics: Precision: 0.856, Recall: 0.832, mAP50: 0.882. YOLOv9 comprehensive metrics, research aims robust processes. Additionally, offering future integration robotics, collection phase can further optimize efficiency enable expansion cultivation areas.

Language: Английский

---

Fruits and vegetables preservation based on AI technology: Research progress and application prospects

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 226, P. 109382 - 109382

Published: Aug. 27, 2024

Language: Английский

---

An innovative approach to detecting the freshness of fruits and vegetables through the integration of convolutional neural networks and bidirectional long short-term memory network

Current Research in Food Science, Journal Year: 2024, Volume and Issue: 8, P. 100723 - 100723

Published: Jan. 1, 2024

Fruit and vegetable freshness testing can improve the efficiency of agricultural product management, reduce resource waste economic losses, plays a vital role in increasing added value fruit products. At present, detection mainly relies on manual feature extraction combined with machine learning. However, features has problem poor adaptability, resulting low detection. Although exist some studies that have introduced deep learning methods to automatically learn characterize fruits vegetables cope diversity variability complex scenes. performance these needs be further improved. Based this, this paper proposes novel method fusion different models extract images correlation between various areas image, so as detect more objectively accurately. First, image size dataset is resized meet input requirements model. Then, characterizing are extracted by fused Finally, parameters model were optimized based model, was evaluated. Experimental results show CNN_BiLSTM which convolutional neural network (CNN) bidirectional long-short term memory (BiLSTM), parameter optimization processing achieve an accuracy 97.76% detecting vegetables. The research promising

Language: Английский

---

FruVeg_MultiNet: A hybrid deep learning-enabled IoT system for fresh fruit and vegetable identification with web interface and customized blind glasses for visually impaired individuals

Journal of Agriculture and Food Research, Journal Year: 2025, Volume and Issue: unknown, P. 101623 - 101623

Published: Jan. 1, 2025

Language: Английский

---

Multiscale bioimpedance detection methods and modeling for dynamic non-destructive monitoring of agricultural product quality

Trends in Food Science & Technology, Journal Year: 2025, Volume and Issue: unknown, P. 104888 - 104888

Published: Jan. 1, 2025

Language: Английский

---

Integrating machine learning, optical sensors, and robotics for advanced food quality assessment and food processing

Food Innovation and Advances, Journal Year: 2025, Volume and Issue: 4(1), P. 65 - 72

Published: Jan. 1, 2025

Language: Английский

---

MIRNet_ECA: Multi-scale inverted residual attention network used for classification of ripeness level for dragon fruit

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127019 - 127019

Published: Feb. 1, 2025

Language: Английский

---

Detection of Rotten Fruits Using XGBoost-Based Deep Learning Algorithm with Explainable Artificial Intelligence Models

Süleyman Demirel Üniversitesi Fen Bilimleri Enstitüsü Dergisi, Journal Year: 2025, Volume and Issue: 29(1), P. 124 - 133

Published: April 25, 2025

Abstract: Achieving high accuracy rates in the field of image processing often exceeds limits a single model. Therefore, hybridizing XGBoost and deep learning models is common approach to obtaining more accurate reliable results. Deep are highly capable extracting complex meaningful features from images. However, effectively classify these features, use powerful machine algorithm like can further enhance performance. Hybrid combine best both models, allowing them achieve higher that would not be possible if used individually. High improves model's reliability effectiveness application, thereby preventing misclassification improving overall hybridization essential for better In this paper, after flattening extracted an XGBoost-based model was trained by utilizing decision trees, achieved 98.813% on test data. SHAP XAI LIME were employed explain model, providing visualizations how impacted decisions positively or negatively based their weight values, demonstrating influenced decision-making process.

Language: Английский

---

Large-scale data-driven uniformity analysis and sensory prediction of commercial banana ripening process

Postharvest Biology and Technology, Journal Year: 2024, Volume and Issue: 219, P. 113203 - 113203

Published: Sept. 13, 2024

Language: Английский

---

Enhancing Apple Cultivar Classification Using Multiview Images

Journal of Imaging, Journal Year: 2024, Volume and Issue: 10(4), P. 94 - 94

Published: April 17, 2024

Apple cultivar classification is challenging due to the inter-class similarity and high intra-class variations. Human experts do not rely on single-view features but rather study each viewpoint of apple identify a cultivar, paying close attention various details. Following our previous work, we try establish similar multiview approach for machine-learning (ML)-based in this paper. In studied using one single view. While these results were promising, it also became clear that view alone might contain enough information case many classes or cultivars. Therefore, exploring task next logical step. Multiview nothing new, use state-of-the-art approaches as base. Our goal find best specific what achievable with given methods towards future applying mobile device without need internet connectivity. study, compare an ensemble model two cases where networks: specialization trained all available images assignment combine separate views into image instance. The latter options reflect dataset organization preprocessing allow smaller models terms stored weights number operations than model. We different based custom dataset. show provides result. However, combined shows decrease accuracy by 3% while requiring only 60% memory weights. Thus, simpler enhanced can open trade-off tasks devices.

Language: Английский

---