Article
Agronomy
Shanxin Zhang, Hao Feng, Shaoyu Han, Zhengkai Shi, Haoran Xu, Yang Liu, Haikuan Feng, Chengquan Zhou, Jibo Yue
Summary: Soybean breeders need different varieties for planting at different latitudes, and timely monitoring of soybean breeding line maturity is crucial for soybean harvesting management. A new convolutional neural network (CNN) called DS-SoybeanNet is designed to improve the performance of unmanned aerial vehicle (UAV)-based soybean maturity monitoring. DS-SoybeanNet can extract and utilize both shallow and deep image features.
Article
Chemistry, Analytical
Daegyun Choi, William Bell, Donghoon Kim, Jichul Kim
Summary: Structural cracks are crucial in assessing the health of aging structures. This study proposes a framework for detecting and locating cracks using image data from a UAV and a deep learning model, showcasing an effective way to identify cracks and their positions.
Article
Environmental Sciences
Milan Bajic, Bozidar Potocnik
Summary: After the military conflict in Ukraine in 2014, several promising solutions for thermal imaging Unexploded Ordnance (UXO) detection were proposed. However, most of the current landmine clearance protocols and practices are still based on outdated technologies. This research introduces an automated UXO detection method and publishes thermal imaging data to bridge these gaps.
Article
Chemistry, Multidisciplinary
Rajagopalan-Sam Rajadurai, Su-Tae Kang
Summary: This study utilized deep convolutional neural networks and transfer learning to detect cracks, achieving a high accuracy rate of 99.9% with the trained model. By fine-tuning the architecture and augmenting the image datasets, the model successfully achieved precise detection and classification of cracks.
APPLIED SCIENCES-BASEL
(2021)
Article
Chemistry, Multidisciplinary
Jiri Maslan, Ludek Cicmanec
Summary: This study explores the automatic detection and evaluation of distress on an airport pavement using unmanned aerial vehicle imagery and artificial intelligence. The YOLOv2 object detector is used for crack detection and the obtained features are processed for position determination and dimension measurement. The study successfully verifies the system on the experimental section of the runway, demonstrating the efficiency and impressive results of unmanned aerial vehicle imagery combined with artificial intelligence.
APPLIED SCIENCES-BASEL
(2023)
Article
Agronomy
Ryoya Tanabe, Tsutomu Matsui, Takashi S. T. Tanaka
Summary: An inexpensive and precise crop yield prediction technology is required for small-scale fields in Asian countries. The effectiveness of convolutional neural networks (CNNs) for crop yield prediction was verified using UAV-based multispectral imagery. The results showed that the CNN model of the heading stage had the lowest RMSE among the four growth stages and outperformed the best linear regression model. These findings suggest that CNN has the potential to improve accuracy, and the heading stage is a suitable data acquisition time for winter wheat.
FIELD CROPS RESEARCH
(2023)
Article
Biology
Xinxin Yang, Mark Stamp
Summary: Low grade endometrial stromal sarcoma (LGESS) is a rare type of uterine cancer, and classic machine learning and deep learning models can be used to assist in its diagnosis. The research shows that deep learning models have a slightly higher classification accuracy compared to classic techniques.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Review
Computer Science, Artificial Intelligence
Abdelmalek Bouguettaya, Hafed Zarzour, Ahmed Kechida, Amine Mohammed Taberkit
Summary: Unmanned Aerial Vehicles (UAVs) have been widely used in agriculture to increase productivity and reduce costs. They can quickly cover large areas and collect valuable data for precision agriculture applications, such as crop classification. Deep learning algorithms, particularly Convolutional Neural Networks (CNN), have become powerful tools for processing UAV-based remote sensing images and improving crop classification accuracy.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Agriculture, Multidisciplinary
Zichen Zhang, Sami Khanal, Amy Raudenbush, Kelley Tilmon, Christopher Stewart
Summary: Severe crop defoliation caused by insects and pests is harmful to agriculture productivity, and machine learning techniques are assessed for their effectiveness in detecting defoliation. Different techniques vary in their ability to accurately characterize defoliation images and avoid misidentifying healthy crops as damaged, with the development of DefoNet, a convolutional neural network designed for detecting crop defoliation, showing potential for practical application in mitigating pest-related crop losses.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2022)
Article
Energy & Fuels
Naveen Venkatesh Sridharan, Vaithiyanathan Sugumaran
Summary: This study proposes an automatic fault classification method for photovoltaic modules using deep learning techniques. The study utilizes several renowned deep convolution neural network models and proposes a hybrid deep ensemble model. Experimental results show that the hybrid model achieves a classification accuracy of 99.04%. Using this technique, faults in photovoltaic modules can be accurately identified, leading to the elimination of downtime and fire hazards.
ENERGY SOURCES PART A-RECOVERY UTILIZATION AND ENVIRONMENTAL EFFECTS
(2022)
Article
Remote Sensing
Mohammed A. Alanezi, Abdullahi Mohammad, Yusuf A. Sha'aban, Houssem R. E. H. Bouchekara, Mohammad S. Shahriar
Summary: The advancement in computing and telecommunication has expanded the applications of drones, including in agriculture. Livestock farming using unmanned aerial vehicle (UAV) systems requires surveillance and monitoring of animals, which necessitates a reliable communication system between UAVs and the ground control station (GCS). The paper proposes learning-based communication strategies and techniques to enable interaction and data exchange between UAVs and a GCS, with the use of a deep auto-encoder UAV design framework for end-to-end communications. Simulation results demonstrate the effectiveness of the auto-encoder in learning joint transmitter and receiver mapping functions for various communication strategies.
Article
Geography, Physical
K. R. Akshatha, A. K. Karunakar, B. Satish Shenoy, K. Phani Pavan, Chinmay V. Dhareshwar, Dennis George Johnson
Summary: Intelligent UAV video analysis has gained attention for its potential in computer vision applications. In order to address the challenge of small object detection, a Manipal-UAV person detection dataset was created, consisting of images captured from UAVs in varying conditions. The dataset provides a benchmark for evaluating state-of-the-art object detection algorithms on small person objects in aerial view scenarios. The dataset is publicly available for researchers to advance UAV and small object detection research.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2023)
Article
Automation & Control Systems
Yuanda Wang, Wenzhang Liu, Jian Liu, Changyin Sun
Summary: This paper investigates the visual navigation and control of a cooperative USV-UAV system for marine search and rescue. A deep learning-based visual detection architecture is developed to extract positional information from UAV images, improving accuracy and efficiency. A reinforcement learning-based USV control strategy is proposed, which can learn a motion control policy with enhanced disturbance rejection ability. Simulation results show stable and accurate position estimation and satisfactory control ability under wave disturbances.
Article
Environmental Sciences
Peng Yang, Kamran Esmaeili, Sebastian Goodfellow, Juan Carlos Ordonez Calderon
Summary: Geological pit wall mapping in surface mining is important for improving geological certainty and operational planning. This study explores the use of drone-acquired RGB images for pit wall mapping. While the results are promising for simple geological settings, they deviate from human-labelled ground truth maps in more complex conditions, highlighting the need for further algorithm optimization for robustness.
Article
Ecology
Brinky Desai, Arpitkumar Patel, Vaishwi Patel, Supan Shah, Mehul S. Raval, Ratna Ghosal
Summary: Individual identification plays a crucial role in studying animal behavior and ecology. This paper presents a non-invasive method using UAV and convolutional neural networks (CNNs) to individually identify free-ranging mugger crocodiles based on their dorsal scute patterns. The trained models showed high efficiency and accuracy, with potential applications in real-world scenarios.
ECOLOGICAL INFORMATICS
(2022)
Article
Construction & Building Technology
Byunghyun Kim, Soojin Cho
STRUCTURAL CONTROL & HEALTH MONITORING
(2019)
Article
Chemistry, Multidisciplinary
Byunghyun Kim, Soojin Cho
APPLIED SCIENCES-BASEL
(2019)
Article
Chemistry, Multidisciplinary
Muhammad Tanveer, Byunghyun Kim, Jonghwa Hong, Sung-Han Sim, Soojin Cho
Summary: This study compared and analyzed the performance of five semantic segmentation models for damage detection in concrete structures, and found that the CGNet model outperformed the others and was effective for on-site damage detection using ECDs.
APPLIED SCIENCES-BASEL
(2022)
Article
Chemistry, Multidisciplinary
Byunghyun Kim, Soojin Cho
APPLIED SCIENCES-BASEL
(2020)