An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images

Object detection is an important task for rapidly localizing target objects using high-resolution satellite imagery (HRSI). Although deep learning has been shown an efficient means of detection, object detection in HRSI remains problematic due to variations in object scale and size. In this article,...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Shenlu Jiang, Wei Yao, Man Sing Wong, Gen Li, Zhonghua Hong, Tae-Yong Kuc, Xiaohua Tong
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2020
Materias:	Artificial intelligence object detection optical image processing Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809
Acceso en línea:	https://doaj.org/article/6acb49be294645a5a5d109f430897a10
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:6acb49be294645a5a5d109f430897a10
record_format	dspace
spelling	oai:doaj.org-article:6acb49be294645a5a5d109f430897a102021-11-13T00:00:19ZAn Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images2151-153510.1109/JSTARS.2020.2975606https://doaj.org/article/6acb49be294645a5a5d109f430897a102020-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9027926/https://doaj.org/toc/2151-1535Object detection is an important task for rapidly localizing target objects using high-resolution satellite imagery (HRSI). Although deep learning has been shown an efficient means of detection, object detection in HRSI remains problematic due to variations in object scale and size. In this article, we present a novel deep neural network (DNN) that combines double-shot neural network with misplaced localization strategy that adapts to object detection tasks in satellite images. This novel architecture optimizes the localization of small and narrow rectangular objects, which frequently appear in HRSI images, without accuracy loss on other size and width/height ratio objects. This method outperforms other state-of-art methods. We evaluated our proposed method on the NWPU VHR-10 public dataset and a new benchmark dataset (seven classes of small and narrow rectangular objects, SNRO-7). The NWPU VHR-10 dataset built a dataset for multiclass object detection; however, most labels are assigned in normal size and width/height ratios. SNRO-7 focuses on multiscale and multisize object detection and includes many small-size and narrow rectangular objects. We also evaluated the accuracy difference on DNN training and testing between gray scale and RGB datasets. The results of the experiment on object detection reveal that the mean average precision (MaP) of our method is 82.6% in NWPU VHR-10 and 79.3% in SNRO-7, which exceeds the MaPs of other state-of-the-art object detection neural networks. The model trained with the RGB dataset can achieve similar accuracy (around 79.0% MIoU) testing in both RGB and gray scale datasets. When training the model by mixing RGB and gray scale datasets in different ratios, the accuracy in the RGB channel significantly decreases with increasing gray scale images, but this does not influence the accuracy in the gray scale dataset.Shenlu JiangWei YaoMan Sing WongGen LiZhonghua HongTae-Yong KucXiaohua TongIEEEarticleArtificial intelligenceobject detectionoptical image processingOcean engineeringTC1501-1800Geophysics. Cosmic physicsQC801-809ENIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol 13, Pp 1068-1081 (2020)
institution	DOAJ
collection	DOAJ
language	EN
topic	Artificial intelligence object detection optical image processing Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809
spellingShingle	Artificial intelligence object detection optical image processing Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809 Shenlu Jiang Wei Yao Man Sing Wong Gen Li Zhonghua Hong Tae-Yong Kuc Xiaohua Tong An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images
description	Object detection is an important task for rapidly localizing target objects using high-resolution satellite imagery (HRSI). Although deep learning has been shown an efficient means of detection, object detection in HRSI remains problematic due to variations in object scale and size. In this article, we present a novel deep neural network (DNN) that combines double-shot neural network with misplaced localization strategy that adapts to object detection tasks in satellite images. This novel architecture optimizes the localization of small and narrow rectangular objects, which frequently appear in HRSI images, without accuracy loss on other size and width/height ratio objects. This method outperforms other state-of-art methods. We evaluated our proposed method on the NWPU VHR-10 public dataset and a new benchmark dataset (seven classes of small and narrow rectangular objects, SNRO-7). The NWPU VHR-10 dataset built a dataset for multiclass object detection; however, most labels are assigned in normal size and width/height ratios. SNRO-7 focuses on multiscale and multisize object detection and includes many small-size and narrow rectangular objects. We also evaluated the accuracy difference on DNN training and testing between gray scale and RGB datasets. The results of the experiment on object detection reveal that the mean average precision (MaP) of our method is 82.6% in NWPU VHR-10 and 79.3% in SNRO-7, which exceeds the MaPs of other state-of-the-art object detection neural networks. The model trained with the RGB dataset can achieve similar accuracy (around 79.0% MIoU) testing in both RGB and gray scale datasets. When training the model by mixing RGB and gray scale datasets in different ratios, the accuracy in the RGB channel significantly decreases with increasing gray scale images, but this does not influence the accuracy in the gray scale dataset.
format	article
author	Shenlu Jiang Wei Yao Man Sing Wong Gen Li Zhonghua Hong Tae-Yong Kuc Xiaohua Tong
author_facet	Shenlu Jiang Wei Yao Man Sing Wong Gen Li Zhonghua Hong Tae-Yong Kuc Xiaohua Tong
author_sort	Shenlu Jiang
title	An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images
title_short	An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images
title_full	An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images
title_fullStr	An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images
title_full_unstemmed	An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images
title_sort	optimized deep neural network detecting small and narrow rectangular objects in google earth images
publisher	IEEE
publishDate	2020
url	https://doaj.org/article/6acb49be294645a5a5d109f430897a10
work_keys_str_mv	AT shenlujiang anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT weiyao anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT mansingwong anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT genli anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT zhonghuahong anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT taeyongkuc anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT xiaohuatong anoptimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT shenlujiang optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT weiyao optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT mansingwong optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT genli optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT zhonghuahong optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT taeyongkuc optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages AT xiaohuatong optimizeddeepneuralnetworkdetectingsmallandnarrowrectangularobjectsingoogleearthimages
_version_	1718430322807275520

An Optimized Deep Neural Network Detecting Small and Narrow Rectangular Objects in Google Earth Images

Ejemplares similares