High‐Throughput Robotic Phenotyping for Quantifying Tomato Disease Severity Enabled by Synthetic Data and Domain‐Adaptive Semantic Segmentation DOI Creative Commons
Tingshu He, Xingjian Li, Zhenghua Zhang

et al.

Journal of Field Robotics, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 13, 2024

ABSTRACT Plant diseases cause an annual global crop loss of 20%–40%, leading to estimated economic losses 30–50 billion dollars. Tomatoes are susceptible more than 200 diseases. Breeding disease‐resistant cultivars is cost‐effective and environmentally sustainable the frequent use pesticides. Traditional breeding methods for disease resistance, relying on direct visual observation measure disease‐related traits, time‐consuming, inaccurate, expensive, require specific knowledge tomato High‐throughput phenotyping essential reduce labor costs, improve measurement accuracy, expedite release new varieties, thereby effectively identifying crops. Precision agriculture efforts have primarily focused detecting individual leaves under controlled laboratory conditions, neglecting assessment severity entire plant in field. To address this, we created a synthetic data set using existing field leaf sets, leveraging game engine minimize additional labeling. Consequently, developed customized unsupervised domain‐adaptive segmentation algorithm that monitors determines based proportion affected areas. The system‐derived percentages show high correlation with manually labeled data, evidenced by coefficient 0.91. Our research demonstrates feasibility ground robots equipped deep‐learning algorithms monitor potentially accelerating automation standardization whole‐plant monitoring tomatoes. This high‐throughput system can also be adapted analyze other crops similar foliar diseases, such as maize, soybeans, cotton.

Language: Английский

Creating synthetic data sets for training of neural networks for automatic catch analysis in fisheries DOI Creative Commons

Jonatan Sjølund Dyrstad,

Elling Ruud Øye

Computers and Electronics in Agriculture, Journal Year: 2025, Volume and Issue: 233, P. 110160 - 110160

Published: March 3, 2025

Language: Английский

Citations

0
A deep learning based visual inspection of small-batch electronic assembly using few-shot-driven synthetic data DOI
Mingxing Jiang, Tingyu Liu,

Songyang Li

et al.

Journal of Intelligent Manufacturing, Journal Year: 2025, Volume and Issue: unknown

Published: April 26, 2025

Language: Английский

Citations

0
Cyber Security Framework for AI-Enabled Robotics and Drone Systems DOI

Muhammad Javed,

Sher Taj, Rahim Khan

et al.

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 231 - 262

Published: Feb. 21, 2025

The smooth out integration of Artificial Intelligence (AI) in drones and robotic systems has dramatically changed industries, from manufacturing to logistics, healthcare, close observation surveillance. This process enabled unprecedented precision, efficiency, for innovation. However, it also introduces unknown before cybersecurity risks, compromising the confidentiality robotics system, integrity, availability critical data. As AI-enabled drone have become increasingly linked, they are weak harm revealing cyber threats. These kinds threats include unauthorized access, not having approval data breaches (failing observe), system failures, potential disruptions prevent progress infrastructure. relevance mention can be harsh, ranging compromised safety financial losses security. chapter provides a comprehensive examination frameworks specifically designed systems.

Language: Английский

Citations

0
A large-scale lychee image parallel classification algorithm based on spark and deep learning DOI
Yiming Xiao, Jianhua Wang, Hongyi Xiong

et al.

Computers and Electronics in Agriculture, Journal Year: 2025, Volume and Issue: 230, P. 109952 - 109952

Published: Jan. 14, 2025

Language: Английский

Citations

0
Advances in Global Remote Sensing Monitoring of Discolored Pine Trees Caused by Pine Wilt Disease: Platforms, Methods, and Future Directions DOI Open Access
Hao Shi,

Liping Chen,

Meixiang Chen

et al.

Forests, Journal Year: 2024, Volume and Issue: 15(12), P. 2147 - 2147

Published: Dec. 5, 2024

Pine wilt disease (PWD), caused by pine wood nematodes, is a major forest that poses serious threat to global resources. Therefore, the prompt identification of PWD-discolored trees crucial for controlling its spread. Currently, remote sensing primary approach monitoring PWD. This study comprehensively reviews advances in It explores platforms and methods used detection trees, evaluates their precision, provides prospects existing problems. Three observations were made from studies: First, unmanned aerial vehicles (UAVs) are dominant platforms, RGB data sources most commonly identifying trees. Second, deep-learning increasingly applied identify Third, early has gained increasing attention. reveals problems associated with acquisition images algorithms. Future research directions include fusion multiple sensors enhance precision obtain an optimal window period. aimed provide technical references scientific foundations comprehensive control

Language: Английский

Citations

1
High‐Throughput Robotic Phenotyping for Quantifying Tomato Disease Severity Enabled by Synthetic Data and Domain‐Adaptive Semantic Segmentation DOI Creative Commons
Tingshu He, Xingjian Li, Zhenghua Zhang

et al.

Journal of Field Robotics, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 13, 2024

ABSTRACT Plant diseases cause an annual global crop loss of 20%–40%, leading to estimated economic losses 30–50 billion dollars. Tomatoes are susceptible more than 200 diseases. Breeding disease‐resistant cultivars is cost‐effective and environmentally sustainable the frequent use pesticides. Traditional breeding methods for disease resistance, relying on direct visual observation measure disease‐related traits, time‐consuming, inaccurate, expensive, require specific knowledge tomato High‐throughput phenotyping essential reduce labor costs, improve measurement accuracy, expedite release new varieties, thereby effectively identifying crops. Precision agriculture efforts have primarily focused detecting individual leaves under controlled laboratory conditions, neglecting assessment severity entire plant in field. To address this, we created a synthetic data set using existing field leaf sets, leveraging game engine minimize additional labeling. Consequently, developed customized unsupervised domain‐adaptive segmentation algorithm that monitors determines based proportion affected areas. The system‐derived percentages show high correlation with manually labeled data, evidenced by coefficient 0.91. Our research demonstrates feasibility ground robots equipped deep‐learning algorithms monitor potentially accelerating automation standardization whole‐plant monitoring tomatoes. This high‐throughput system can also be adapted analyze other crops similar foliar diseases, such as maize, soybeans, cotton.

Language: Английский

Citations

0