arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2010.08720cs.CV
收藏

PolarDet:一种快速、高精度的航空图像旋转目标检测器

PolarDet: A Fast, More Precise Detector for Rotated Target in Aerial Images

Harbin Institute of Technology(哈尔滨工业大学);Hikvision Research Institute(海康威视研究院)

Pengbo Zhao, Zhenshen Qu, Yingjia Bu, Wenming Tan, Ye Ren, Shiliang Pu

摘要:

高分辨率航空影像的快速、精确目标检测一直是一项具有挑战性的任务。由于对象尺度、旋转和长宽比的急剧变化,大多数现有方法效率低下且不精确。在本文中,我们表示的极坐标方法在极坐标方向的对象,并提出PolarDet,一个快速,准确的一级目标检测器的基础上表示。我们的检测器引入了一个亚像素中心的语义结构,以进一步提高分类的准确性。PolarDet在空中目标检测任务中实现了几乎所有的SOTA性能,推理速度更快。具体地说,我们的方法在DOTA、UCAS-AOD和HRSC上分别获得了76.64mAP、97.01mAP和90.46mAP的SOTA结果。最值得注意的是,我们的PolarDet在UCAS-AOD数据集上获得了最佳性能,并达到了最快的速度(32 fps)。

英文摘要:

Fast and precise object detection for high-resolution aerial images has been a challenging task over the years. Due to the sharp variations on object scale, rotation, and aspect ratio, most existing methods are inefficient and imprecise. In this paper, we represent the oriented objects by polar method in polar coordinate and propose PolarDet, a fast and accurate one-stage object detector based on that representation. Our detector introduces a sub-pixel center semantic structure to further improve classifying veracity. PolarDet achieves nearly all SOTA performance in aerial object detection tasks with faster inference speed. In detail, our approach obtains the SOTA results on DOTA, UCAS-AOD, HRSC with 76.64\% mAP, 97.01\% mAP, and 90.46\% mAP respectively. Most noticeably, our PolarDet gets the best performance and reaches the fastest speed(32fps) at the UCAS-AOD dataset.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑