An Efficient and Robust Star Identification Algorithm Based on Neural Networks

A lost-in-space star identification algorithm based on a one-dimensional Convolutional Neural Network (1D CNN) is proposed. The lost-in-space star identification aims to identify stars observed with corresponding catalog stars when there is no prior attitude information. With the help of neural netw...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Bendong Wang, Hao Wang, Zhonghe Jin
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	star identification modified log-polar mapping one-dimensional Convolutional NeuralNetwork Chemical technology TP1-1185
Acceso en línea:	https://doaj.org/article/efd8c16042194ebcb92e56406e4c12ee
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:efd8c16042194ebcb92e56406e4c12ee
record_format	dspace
spelling	oai:doaj.org-article:efd8c16042194ebcb92e56406e4c12ee2021-11-25T18:58:29ZAn Efficient and Robust Star Identification Algorithm Based on Neural Networks10.3390/s212276861424-8220https://doaj.org/article/efd8c16042194ebcb92e56406e4c12ee2021-11-01T00:00:00Zhttps://www.mdpi.com/1424-8220/21/22/7686https://doaj.org/toc/1424-8220A lost-in-space star identification algorithm based on a one-dimensional Convolutional Neural Network (1D CNN) is proposed. The lost-in-space star identification aims to identify stars observed with corresponding catalog stars when there is no prior attitude information. With the help of neural networks, the robustness and the speed of the star identification are improved greatly. In this paper, a modified log-Polar mapping is used to constructed rotation-invariant star patterns. Then a 1D CNN is utilized to classify the star patterns associated with guide stars. In the 1D CNN model, a global average pooling layer is used to replace fully-connected layers to reduce the number of parameters and the risk of overfitting. Experiments show that the proposed algorithm is highly robust to position noise, magnitude noise, and false stars. The identification accuracy is <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>98.1</mn><mo>%</mo></mrow></semantics></math></inline-formula> with 5 pixels position noise, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>97.4</mn><mo>%</mo></mrow></semantics></math></inline-formula> with 5 false stars, and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>97.7</mn><mo>%</mo></mrow></semantics></math></inline-formula> with 0.5 Mv magnitude noise, respectively, which is significantly higher than the identification rate of the pyramid, optimized grid and modified log-polar algorithms. Moreover, the proposed algorithm guarantees a reliable star identification under dynamic conditions. The identification accuracy is <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>82.1</mn><mo>%</mo></mrow></semantics></math></inline-formula> with angular velocity of 10 degrees per second. Furthermore, its identification time is as short as 32.7 miliseconds and the memory required is about 1920 kilobytes. The algorithm proposed is suitable for current embedded systems.Bendong WangHao WangZhonghe JinMDPI AGarticlestar identificationmodified log-polar mappingone-dimensional Convolutional NeuralNetworkChemical technologyTP1-1185ENSensors, Vol 21, Iss 7686, p 7686 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	star identification modified log-polar mapping one-dimensional Convolutional NeuralNetwork Chemical technology TP1-1185
spellingShingle	star identification modified log-polar mapping one-dimensional Convolutional NeuralNetwork Chemical technology TP1-1185 Bendong Wang Hao Wang Zhonghe Jin An Efficient and Robust Star Identification Algorithm Based on Neural Networks
description	A lost-in-space star identification algorithm based on a one-dimensional Convolutional Neural Network (1D CNN) is proposed. The lost-in-space star identification aims to identify stars observed with corresponding catalog stars when there is no prior attitude information. With the help of neural networks, the robustness and the speed of the star identification are improved greatly. In this paper, a modified log-Polar mapping is used to constructed rotation-invariant star patterns. Then a 1D CNN is utilized to classify the star patterns associated with guide stars. In the 1D CNN model, a global average pooling layer is used to replace fully-connected layers to reduce the number of parameters and the risk of overfitting. Experiments show that the proposed algorithm is highly robust to position noise, magnitude noise, and false stars. The identification accuracy is <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>98.1</mn><mo>%</mo></mrow></semantics></math></inline-formula> with 5 pixels position noise, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>97.4</mn><mo>%</mo></mrow></semantics></math></inline-formula> with 5 false stars, and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>97.7</mn><mo>%</mo></mrow></semantics></math></inline-formula> with 0.5 Mv magnitude noise, respectively, which is significantly higher than the identification rate of the pyramid, optimized grid and modified log-polar algorithms. Moreover, the proposed algorithm guarantees a reliable star identification under dynamic conditions. The identification accuracy is <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>82.1</mn><mo>%</mo></mrow></semantics></math></inline-formula> with angular velocity of 10 degrees per second. Furthermore, its identification time is as short as 32.7 miliseconds and the memory required is about 1920 kilobytes. The algorithm proposed is suitable for current embedded systems.
format	article
author	Bendong Wang Hao Wang Zhonghe Jin
author_facet	Bendong Wang Hao Wang Zhonghe Jin
author_sort	Bendong Wang
title	An Efficient and Robust Star Identification Algorithm Based on Neural Networks
title_short	An Efficient and Robust Star Identification Algorithm Based on Neural Networks
title_full	An Efficient and Robust Star Identification Algorithm Based on Neural Networks
title_fullStr	An Efficient and Robust Star Identification Algorithm Based on Neural Networks
title_full_unstemmed	An Efficient and Robust Star Identification Algorithm Based on Neural Networks
title_sort	efficient and robust star identification algorithm based on neural networks
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/efd8c16042194ebcb92e56406e4c12ee
work_keys_str_mv	AT bendongwang anefficientandrobuststaridentificationalgorithmbasedonneuralnetworks AT haowang anefficientandrobuststaridentificationalgorithmbasedonneuralnetworks AT zhonghejin anefficientandrobuststaridentificationalgorithmbasedonneuralnetworks AT bendongwang efficientandrobuststaridentificationalgorithmbasedonneuralnetworks AT haowang efficientandrobuststaridentificationalgorithmbasedonneuralnetworks AT zhonghejin efficientandrobuststaridentificationalgorithmbasedonneuralnetworks
_version_	1718410499865968640

An Efficient and Robust Star Identification Algorithm Based on Neural Networks

Ejemplares similares