Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection

The vehicle detection in remote sensing images is a challenging task due to the small size of the objects and interference of a complex background. Traditional methods require a large number of anchor boxes, and the intersection rate between these anchor boxes and an object’s real position boxes nee...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Lijian Yu, Xiyang Zhi, Jianming Hu, Shikai Jiang, Wei Zhang, Wenbin Chen
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	vehicle detection keypoint detection Gaussian label deep learning remote sensing images Science Q
Acceso en línea:	https://doaj.org/article/f020922838cb4bd5938d8873eb6c028f
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:f020922838cb4bd5938d8873eb6c028f
record_format	dspace
spelling	oai:doaj.org-article:f020922838cb4bd5938d8873eb6c028f2021-11-11T18:56:50ZSmall-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection10.3390/rs132144422072-4292https://doaj.org/article/f020922838cb4bd5938d8873eb6c028f2021-11-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/21/4442https://doaj.org/toc/2072-4292The vehicle detection in remote sensing images is a challenging task due to the small size of the objects and interference of a complex background. Traditional methods require a large number of anchor boxes, and the intersection rate between these anchor boxes and an object’s real position boxes needs to be high enough. Moreover, the size and aspect ratio of each anchor box need to be designed manually. For small objects, more anchor boxes need to be set. To solve these problems, we regard the small object as a keypoint in the relevant background and propose an anchor-free vehicle detection network (AVD-kpNet) to robustly detect small-sized vehicles in remote sensing images. The AVD-kpNet framework fuses features across layers with a deep layer aggregation architecture, preserving the fine features of small objects. First, considering the correlation between the object and the surrounding background, a 2D Gaussian distribution strategy is adopted to describe the ground truth, instead of a hard label approach. Moreover, we redesign the corresponding focus loss function. Experimental results demonstrate that our method has a higher accuracy for the small-sized vehicle detection task in remote sensing images compared with several advanced methods.Lijian YuXiyang ZhiJianming HuShikai JiangWei ZhangWenbin ChenMDPI AGarticlevehicle detectionkeypoint detectionGaussian labeldeep learningremote sensing imagesScienceQENRemote Sensing, Vol 13, Iss 4442, p 4442 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	vehicle detection keypoint detection Gaussian label deep learning remote sensing images Science Q
spellingShingle	vehicle detection keypoint detection Gaussian label deep learning remote sensing images Science Q Lijian Yu Xiyang Zhi Jianming Hu Shikai Jiang Wei Zhang Wenbin Chen Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection
description	The vehicle detection in remote sensing images is a challenging task due to the small size of the objects and interference of a complex background. Traditional methods require a large number of anchor boxes, and the intersection rate between these anchor boxes and an object’s real position boxes needs to be high enough. Moreover, the size and aspect ratio of each anchor box need to be designed manually. For small objects, more anchor boxes need to be set. To solve these problems, we regard the small object as a keypoint in the relevant background and propose an anchor-free vehicle detection network (AVD-kpNet) to robustly detect small-sized vehicles in remote sensing images. The AVD-kpNet framework fuses features across layers with a deep layer aggregation architecture, preserving the fine features of small objects. First, considering the correlation between the object and the surrounding background, a 2D Gaussian distribution strategy is adopted to describe the ground truth, instead of a hard label approach. Moreover, we redesign the corresponding focus loss function. Experimental results demonstrate that our method has a higher accuracy for the small-sized vehicle detection task in remote sensing images compared with several advanced methods.
format	article
author	Lijian Yu Xiyang Zhi Jianming Hu Shikai Jiang Wei Zhang Wenbin Chen
author_facet	Lijian Yu Xiyang Zhi Jianming Hu Shikai Jiang Wei Zhang Wenbin Chen
author_sort	Lijian Yu
title	Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection
title_short	Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection
title_full	Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection
title_fullStr	Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection
title_full_unstemmed	Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection
title_sort	small-sized vehicle detection in remote sensing image based on keypoint detection
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/f020922838cb4bd5938d8873eb6c028f
work_keys_str_mv	AT lijianyu smallsizedvehicledetectioninremotesensingimagebasedonkeypointdetection AT xiyangzhi smallsizedvehicledetectioninremotesensingimagebasedonkeypointdetection AT jianminghu smallsizedvehicledetectioninremotesensingimagebasedonkeypointdetection AT shikaijiang smallsizedvehicledetectioninremotesensingimagebasedonkeypointdetection AT weizhang smallsizedvehicledetectioninremotesensingimagebasedonkeypointdetection AT wenbinchen smallsizedvehicledetectioninremotesensingimagebasedonkeypointdetection
_version_	1718431679873286144

Small-Sized Vehicle Detection in Remote Sensing Image Based on Keypoint Detection

Ejemplares similares