SIVED: a SAR Image dataset for VEhicle Detection

Introduction

SIVED is a SAR image dataset for vehicle detection using Ka, Ku, and X bands of data. Rotatable bounding box annotations were employed to improve positioning accuracy.

Baic Information

• Raw Data

  Data	Source	Band	Polarization	Resolution
  FARAD	Sandia National Laboratory	Ka/X	VV/HH	0.1m×0.1m
  MiniSAR	Sandia National Laboratory	Ku	-	0.1m×0.1m
  MSTAR	U.S. Air Force	X	HH	0.3m×0.3m

• Statistics

  	Scene	Train	Valid	Test	Total
  number of chips	urban	578	72	71	721	1044
  	MSTAR	259	32	32	323
  number of vehicles	urban	5417	710	718	6845	12013
  	MSTAR	4144	512	512	5168

• Annotation

  XML (reference PASCAL VOC) and TXT (reference DOTA)

• File Structure

  

  Cite

  If you feel the dataset is useful, please cite as the following format.

  @Article{rs15112825,

  AUTHOR = {Lin, Xin and Zhang, Bo and Wu, Fan and Wang, Chao and Yang, Yali and Chen, Huiqin},

  TITLE = {SIVED: A SAR Image Dataset for Vehicle Detection Based on Rotatable Bounding Box},

  JOURNAL = {Remote Sensing},

  VOLUME = {15},

  YEAR = {2023},

  NUMBER = {11},

  ARTICLE-NUMBER = {2825},

  URL = { https://www.mdpi.com/2072-4292/15/11/2825 },

  ISSN = {2072-4292},

  DOI = {10.3390/rs15112825}

  }