Related references
Note: Only part of the references are listed.
Article
Environmental Sciences
Dahang Wan et al.
Summary: This paper proposes an enhanced YOLOv5 algorithm for object detection in high-resolution optical remote sensing images, which uses multiple layers of the feature pyramid, a multi-detection-head strategy, and a hybrid attention module to improve the effect of object-detection networks.
Article
Geosciences, Multidisciplinary
Kushanav Bhuyan et al.
Summary: The study successfully detected and characterized building footprints using deep learning and auxiliary data, creating an elements-at-risk (EaR) database. The method was applied to different cities in Kerala, India, with high accuracy. The research contributes to regional risk assessment, disaster risk mitigation, and policy development.
Article
Remote Sensing
Lei Yu et al.
Summary: Synthetic Aperture Radar (SAR) is vital for marine ship management, but the complex environment and imaging limitations result in significant noise interference in SAR imaging, especially for ship detection algorithms. To overcome this challenge, we propose a detection algorithm based on convolutional neural network and transformer, which shows improved performance in detecting ships in noise-laden SAR images compared to other deep learning models.
REMOTE SENSING LETTERS
(2023)
Article
Environmental Sciences
Zhuo Chen et al.
Summary: This paper proposes a complex scenes multi-scale ship detection model, called CSD-YOLO, which improves detection accuracy and the model's capacity by using the SAS-FPN module and SIoU loss function. Thorough tests on the HRSID and SSDD datasets show that CSD-YOLO achieves better detection performance than the baseline YOLOv7 and other deep learning-based methods.
Article
Chemistry, Multidisciplinary
Bowen Sun et al.
Summary: This study proposes a detection network consisting of three stages (preattention, attention, and prediction) to address the problem of existing detection networks in neglecting global aspects and lacking automatic adjustment based on input characteristics. The proposed model achieves the highest accuracy compared to other state-of-the-art models on the SAR Ship Detection dataset.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Fang Xiaolin et al.
Summary: Accurate objects detection in remote sensing images is crucial for security, transportation, and rescue applications. This paper proposes an improved model to address the challenge of detecting small-sized objects in remote sensing images, achieving higher accuracy in experiments.
PATTERN RECOGNITION LETTERS
(2022)
Article
Environmental Sciences
Jiahang Liu et al.
Summary: This article proposes an object detection algorithm for remote sensing images based on multi-receptive-field features and relation-connected attention. The algorithm utilizes dilated convolution to aggregate context information from different receptive fields, allowing for better adaptation to object scale changes in complex scenarios. The inclusion of a relation-connected attention module, which combines both global and local attention, enhances feature discriminability and detector robustness. Experimental results demonstrate that these modules effectively improve the performance of basic deep CNNs and achieve better results in multi-scale object detection in complex backgrounds.
Review
Computer Science, Information Systems
Jaskirat Kaur et al.
Summary: This study provides a detailed literature review on object detection techniques and discusses their differences and performance evaluation. It also explores dataset preparation, annotation tools, and future directions for object detection research. Comparative analysis reveals variations in architecture, optimization function, and training strategies among different object detection techniques, with deep neural networks showing notable success in improving detector performance.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Review
Environmental Sciences
Zheng Li et al.
Summary: This paper reviews the development history of remote sensing object detection techniques and systematically summarizes the steps used in deep learning-based detection algorithms. It introduces a taxonomy based on various detection methods, summarizing major improvement strategies such as attention mechanisms, multi-scale feature fusion, and super-resolution. It also presents recognized open-source benchmarks and evaluation metrics for remote sensing detection. Lastly, it discusses the challenges and potential trends in the field of RSOD.
Review
Plant Sciences
Saqib Ali Nawaz et al.
Summary: This paper introduces the importance of object detection in machine vision and deep learning, and the recent progress of object detection algorithms based on deep learning in aspects such as data processing, network structure, and loss function. The latest improvement ideas of typical object detection algorithms are discussed, and future research directions are surveyed.
FRONTIERS IN PLANT SCIENCE
(2022)
Article
Environmental Sciences
Yan Wang et al.
Summary: In this paper, we propose a novel context information refinement few-shot detector (CIR-FSD) for remote sensing images. By designing a context information refinement module and improving the region proposal network, discriminant context features can be effectively extracted and more positive anchors for novel categories can be obtained.
Article
Computer Science, Information Systems
Byungjin Ku et al.
Summary: Wearing a hard hat can significantly improve the safety of workers on construction sites. However, workers often remove their helmets due to a weak sense of safety and discomfort, which poses a significant danger. Manual monitoring of helmet wearing is labor-intensive and challenging to implement widely. In this paper, an AI method is proposed to detect helmet wearing with high accuracy. The proposed method combines YOLOv4 with an image super resolution module and utilizes dense blocks to reduce computation and network structure parameters. Experimental results show that the proposed method outperforms currently available small target detection methods and achieves an average precision of 93.3%.
Article
Chemistry, Analytical
Lei Pang et al.
Summary: Unlike optical satellites, SAR satellites can work all day and in all weather conditions, making them useful for ocean monitoring. Ship monitoring accuracy issues can be solved using a lightweight YOLOV5-MNE method, which improves detection performance through module redesign and attention mechanism.
Proceedings Paper
Engineering, Marine
Athira Nambiar et al.
Summary: This paper proposes a method for improving ship detection in Synthetic Aperture Radar (SAR) images using advanced deep learning techniques. Various benchmark models are compared to evaluate their detection performance in publicly available SAR datasets, and ship lateral images are detected in real-time scenarios. The experimental results demonstrate the effectiveness of deep learning models in ship detection and tracking applications.
2022 OCEANS HAMPTON ROADS
(2022)
Article
Imaging Science & Photographic Technology
Krishna Patel et al.
Summary: The remote sensing surveillance of maritime areas is crucial for both security and environmental reasons. Machine learning methods, particularly deep learning with convolutional neural networks, have shown effectiveness in automatic ship classification from satellite images. This paper focuses on the development and comparison of different versions of the YOLO algorithm for ship detection, finding that YOLOv5 outperforms YOLOv4 and YOLOv3 in terms of accuracy.
JOURNAL OF IMAGING
(2022)
Article
Chemistry, Analytical
Zhiqiang Zhang et al.
Summary: SEFPN is a new network structure proposed in this article, which improves the overall network performance by balancing the semantic representation of each layer of features.
Article
Ratna Kumari Vemuri et al.
Arabian Journal of Geosciences
(2021)
Article
Chemistry, Analytical
Lei Lang et al.
Summary: The study introduces a lightweight object detection method for remote sensing images, achieving high-speed and high-accuracy detection through efficient channel attention layers and differential evolution algorithm. Experimental results show that the network outperforms existing lightweight models in accuracy and performs well on embedded boards.
Review
Remote Sensing
Sabelo P. Simelane et al.
Summary: This paper summarizes the progress made in the Kingdom of Eswatini regarding the application of remote sensing and GIS in LULC monitoring and classification, highlighting some shortcomings in covered thematic areas, image classification methods, and approaches for improving classification accuracy.
SOUTH AFRICAN JOURNAL OF GEOMATICS
(2021)
Article
Geography, Physical
Ke Li et al.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2020)
Article
Environmental Sciences
Xin Zhang et al.
Review
Computer Science, Artificial Intelligence
Kang Tong et al.
IMAGE AND VISION COMPUTING
(2020)
Article
Environmental Sciences
Minh-Tan Pham et al.
Article
Chemistry, Analytical
Min Li et al.
Article
Environmental Sciences
Fei Gao et al.
Article
Multidisciplinary Sciences
Ray- Chang et al.
Proceedings Paper
Computer Science, Artificial Intelligence
Vung Pham et al.
2020 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA)
(2020)
Review
Computer Science, Theory & Methods
Shahid Karim et al.
MULTIAGENT AND GRID SYSTEMS
(2020)
Article
Environmental Sciences
Lalit Kumar et al.
Proceedings Paper
Computer Science, Artificial Intelligence
Bharat Singh et al.
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2018)
Article
Geochemistry & Geophysics
Yang Long et al.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2017)