Integrating deep learning with non‐destructive thermal imaging for precision guava ripeness determination

Journal of the Science of Food and Agriculture, Journal Year: 2024, Volume and Issue: 104(13), P. 7843 - 7853

Published: May 28, 2024

Abstract BACKGROUND To mitigate post‐harvest losses and inform harvesting decisions at the same time as ensuring fruit quality, precise ripeness determination is essential. The complexity arises in assessing guava a result of subtle alterations some varieties during ripening process, making visual assessment less reliable. present study proposes non‐destructive method employing thermal imaging for assessment, involving obtaining images samples different stages, followed by data pre‐processing. Five deep learning models (AlexNet, Inception‐v3, GoogLeNet, ResNet‐50 VGGNet‐16) were applied, their performances systematically evaluated compared. RESULTS VGGNet‐16 demonstrated outstanding performance, achieving average precision 0.92, sensitivity 0.93, specificity 0.96, F1‐score 0.92 accuracy within training duration 484 s. CONCLUSION presents scalable approach determination, contributing to waste reduction enhancing efficiency supply chains production. These initiatives align with environmentally friendly practices agriculture. © 2024 Society Chemical Industry.

Language: Английский

---

Vision-Based Tactile Sensors in Precision Agriculture: Deep Learning Approaches, Applications, and Limitations

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: April 9, 2024

Abstract The integration of artificial intelligence with sensor technologies has revolutionized precision agriculture, offering unprecedented opportunities for enhancing crop management and productivity. This review focuses on the latest advancements in vision-based tactile sensors, a technology at forefront this transformation. By combining data techniques, these sensors provide more comprehensive understanding agricultural environment. We investigate thoroughly role deep learning approaches refining functionality highlighting their potential to significantly improve accuracy efficiency operations. paper also explores importance specialized datasets training neural networks applications, assessing current landscape identifying gaps available data. Through thorough examination state art, aims shed light AI-driven sensing agriculture outline future research directions further advance field.

Language: Английский

---

An improved YOLOv8n-IRP model for natural rubber tree tapping surface detection and tapping key point positioning

Frontiers in Plant Science, Journal Year: 2024, Volume and Issue: 15

Published: Oct. 30, 2024

Aiming at the problem that lightweight algorithm models are difficult to accurately detect and locate tapping surfaces key points in complex rubber forest environments, this paper proposes an improved YOLOv8n-IRP model based on YOLOv8n-Pose. First, receptive field attention mechanism is introduced into backbone network enhance feature extraction ability of surface. Secondly, AFPN structure used reduce loss degradation low-level high-level information. Finally, designs a dual-branch point detection head improve screening features In performance comparison experiment, improves D_mAP50 P_mAP50 by 1.4% 2.3%, respectively, over original while achieving average success rate 87% variable illumination test, which demonstrates enhanced robustness. positioning achieves overall better localization than YOLOv8n-Pose YOLOv5n-Pose, realizing Euclidean distance error less 40 pixels. summary, shows excellent performance, not only provides new method for rubber-tapping robot but also technical support unmanned operation intelligent robot.

Language: Английский

---

A multi-source feature stable learning method for rapid identification of cork spot disorder in ‘Akizuki’ pear

Postharvest Biology and Technology, Journal Year: 2024, Volume and Issue: 219, P. 113285 - 113285

Published: Oct. 30, 2024

Language: Английский

---