NDFTC: A New Detection Framework of Tropical Cyclones from Meteorological Satellite Images with Deep Transfer Learning

Pang, Shanchen; Xie, Pengfei; Xu, Danya; Fan, Meng; Tao, Xixi; Li, Bowen; Liu, Ying; Song, Tao

doi:10.3390/rs13091860

Cited by 27 publications

(16 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deo et al (2017) assessed the relationship between different types of cyclones by using transfer learning and traditional neural network methods to achieve more stable intensity predictions for tropical cyclones. Pang et al (2021) combined a deep convolutional generative adversarial network (DCGAN) and the YOLOv3 model to propose a New Detection Framework of Tropical Cyclones (NDFTC) with good stability and accuracy. Combinido et al (2018).…”

Section: Transfer Learning-based Methodsmentioning

confidence: 99%

USFP: An unbalanced severe typhoon formation prediction framework based on transfer learning

et al. 2023

View full text Add to dashboard Cite

IntroductionSevere typhoons, as extreme weather events, can cause a large number of casualties and property damage in coastal areas. There are mainly three kinds of methods for the prediction of severe typhoon formation, which are the numerical-based methods, the statistical-based methods, and the machine learning-based methods. However, existing methods do not consider the unbalance between the number of ordinary typhoon samples and severe typhoon samples, which makes the accuracies of existing methods in the prediction of severe typhoons much lower than that of ordinary typhoons.MethodsIn this paper, we propose an unbalanced severe typhoon formation prediction (USFP) framework based on transfer learning. We first propose a severe typhoon pre-learning model which is used to learn prior knowledge from a constructed balanced dataset. Then, we propose an unbalanced severe typhoon re-learning model which utilizes the prior knowledge learning from the pre-learning model. Our USFP framework fuses three different variables, which are atmospheric variables, sea surface variables, and ocean hydrographic variables.ResultsExtensive experiments based on datasets of three different regions show that our USFP framework outperforms the numerical model IFS of ECMWF and existing machine learning methods.

show abstract

Section: Transfer Learning-based Methodsmentioning

confidence: 99%

USFP: An unbalanced severe typhoon formation prediction framework based on transfer learning

et al. 2023

View full text Add to dashboard Cite

show abstract

“…In order to mitigate the disasters of tropical cyclones and reduce economic losses, the path prediction and forecast of tropical cyclones [1,2] are very important. Meteorological satellites provide reliable solutions for observing tropical cyclones [3,4]. According to satellite images, the path of tropical cyclone can be analyzed and its scope can be predicted [5].…”

Section: Introductionmentioning

confidence: 99%

A method of locating tropical cyclone center based on gaussian model

You

Qian

2023

Preprint

View full text Add to dashboard Cite

The Tropical cyclones usually refer to low-pressure eddies that rotate rapidly in the tropics. It seriously threatens the life and property safety,Therefore, the prediction of both the accurate location of the tropical cyclone center and the tropical cyclone path is the key to preventing tropical cyclone disasters. In this paper, a fast method for calculating the center of tropical cyclone using infrared satellite image or visible satellite image is proposed. A two-dimensional Gaussian normal distribution model is used to simulate the gray characteristics of tropical cyclone images. The gray distribution of tropical cyclone satellite image approximates to a two-dimensional Gaussian distribution. In multiple directions, the pixel gray values of tropical cyclone are accumulated, and a set of one-dimensional Gaussian normal distribution functions are used to fit these accumulated values. The peak positions of these one-dimensional Gaussian functions are calculated by derivation. Then, a linear function is fitted according to the relationship between the peak position of the group of Gaussian functions and the selected angle. The coefficient of the fitting linear function is exactly the central coordinate of the tropical cyclone. The proposed method for locating the center of tropical cyclone is tested using visible and infrared satellite images. The results are compared with the best trajectory provided by the Japan Meteorological Agency.

show abstract

“…Artificial intelligence has experienced great development in oceanography and meteorology [23][24][25][26][27][28]. Some work has also focused on utilizing artificial intelligence to optimize Delaunay triangulation-based unstructured meshes.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Machine Learning for Improved Delaunay Triangulation

Song

Wang

et al. 2021

JMSE

Self Cite

View full text Add to dashboard Cite

Physical oceanography models rely heavily on grid discretization. It is known that unstructured grids perform well in dealing with boundary fitting problems in complex nearshore regions. However, it is time-consuming to find a set of unstructured grids in specific ocean areas, particularly in the case of land areas that are frequently changed by human construction. In this work, an attempt was made to use machine learning for the optimization of the unstructured triangular meshes formed with Delaunay triangulation in the global ocean field, so that the triangles in the triangular mesh were closer to equilateral triangles, the long, narrow triangles in the triangular mesh were reduced, and the mesh quality was improved. Specifically, we used Delaunay triangulation to generate the unstructured grid, and then developed a K-means clustering-based algorithm to optimize the unstructured grid. With the proposed method, unstructured meshes were generated and optimized for global oceans, small sea areas, and the South China Sea estuary to carry out data experiments. The results suggested that the proportion of triangles with a triangle shape factor greater than 0.7 amounted to 77.80%, 79.78%, and 79.78%, respectively, in the unstructured mesh. Meanwhile, the proportion of long, narrow triangles in the unstructured mesh was decreased to 8.99%, 3.46%, and 4.12%, respectively.

show abstract

NDFTC: A New Detection Framework of Tropical Cyclones from Meteorological Satellite Images with Deep Transfer Learning

Cited by 27 publications

References 50 publications

USFP: An unbalanced severe typhoon formation prediction framework based on transfer learning

USFP: An unbalanced severe typhoon formation prediction framework based on transfer learning

A method of locating tropical cyclone center based on gaussian model

Unsupervised Machine Learning for Improved Delaunay Triangulation

Contact Info

Product

Resources

About