[1]何维娟,江 涛,王林飞,等.舰船遥感图像数据集 DSTD 的构建研究[J].计算机技术与发展,2022,32(07):82-86.[doi:10. 3969 / j. issn. 1673-629X. 2022. 07. 014]
　HE Wei-juan,JIANG Tao,WANG Lin-fei,et al.Research on DSTD Construction of Ship Remote Sensing Image Dataset[J].,2022,32(07):82-86.[doi:10. 3969 / j. issn. 1673-629X. 2022. 07. 014]
点击复制

舰船遥感图像数据集 DSTD 的构建研究()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年07期
页码:
82-86
栏目:
图形与图像
出版日期:
2022-07-10

文章信息/Info

Title:
Research on DSTD Construction of Ship Remote Sensing Image Dataset
文章编号:
1673-629X(2022)07-0082-05
作者:
何维娟1; 江 涛1; 王林飞2; 徐权峰1; 王 欣1
1. 云南民族大学 数学与计算机科学学院,云南 昆明 650500;
2. 云南大学 信息学院,云南 昆明 650091
Author(s):
HE Wei-juan1 ; JIANG Tao1 ; WANG Lin-fei2 ; XU Quan-feng1 ; WANG Xin1
1. School of Mathematics and Computer Science,Yunnan Minzu University,Kunming 650500,China;
2. School of Information,Yunnan University,Kunming 650091,China
关键词:
目标检测; 遥感; 数据集; 舰船; YOLOV5
Keywords:
object detection; remote sensing; dataset; ship; YOLOV5
分类号:
TP7
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 07. 014
摘要:
目标检测算法在 PASCAL VOC、COCO 等一系列数据集中都取得了一定的效果,但是都是面向自然场景下的多目标检测任务,而这些数据集中的单类目标场景往往很单一,相应目标数量也很少,针对特定场景和特定目标的数据集并不多见。 而对于此类任务的数据集往往也是很有价值的,例如遥感场景下的舰船检测或者飞机检测。 针对此问题, 构建一种面向遥感场景的大规模水面舰船目标检测数据集,其数据集的主要来源为 DIOR、DOTA、NWPU - VHRdataset、TGRS -HRRSD-Dataset-master 等几个开源数据集,将其命名为 DSTD( dataset for ship target detection),数据集中包含 4 854 张舰船图片,87 076 个舰船实例。 DSTD 数据集具有数量多、多尺度和成像差异大以及较高的类内多样性等特点。 在构建数据集的基础上,进一步分析了遥感舰船图像的细节特征,评估了一些经典目标检测方法的性能,并进行了实验对比,验证了该数据集的可行性,同时发现了当前较适合舰船的检测方法:即 YOLOV5。 使用 YOLOV5 算法进行舰船图像目标检测,在保证高检测精度的同时,能保持极快的检测速度。
Abstract:
Target detection algorithms have achieved certain effects in a series of datasets such as PASCAL VOC and COCO,but they are all for multi-target detection tasks in natural scenes. The single-class target scenes in these data sets are often single and the number of corresponding targets is small. Datasets for specific scenes and targets are rare. Datasets for such tasks are often valuable,such as ship detection or aircraft detection in remote sensing scenarios. To solve this problem, a large - scale surface ship target detection dataset for remote sensing scene is proposed. The main sources? ? of the dataset are DIOR,DOTA,NWPU - VHRdataset,TGRS - HRRSD - Dataset -master and so on,called DSTD ( dataset for ship target detection) . The dataset contains 4 854 ship pictures and 87 076 ship instances.DSTD are characterized by large number,large multi-scale and imaging differences,and high intra-class diversity. On the basis of the proposed dataset,we further analyze the detailed features of remote sensing ship images,evaluate the performance of some classical target detection methods,and carry out experiments to verify the feasibility of the dataset. At the same time, YOLOV5,a more suitable ship detection method,is found. The YOLOV5 algorithm is used to detect ship image targets, which ensures high detection accuracy while maintaining extremely fast detection speed.

相似文献/References:

[1]刘晓明 李毓蕙 高燕 郑华强.基于目标区域清晰显示的H.264编码策略[J].计算机技术与发展,2010,(06):29.
　LIU Xiao-ming,LI Yu-hui,GAO Yan,et al.A Coding Strategy of H.264 Based on High-definition Display of Target Region[J].,2010,(07):29.
[2]刘翔 吴谨 祝愿博 康晓晶.基于视频序列的目标检测与跟踪技术研究[J].计算机技术与发展,2009,(11):179.
　LIU Xiang,WU Jin,ZHU Yuan-bo,et al.A Study of Object Detecting and Tracking Based on Video Sequences[J].,2009,(07):179.
[3]曙光 张超 蔡则苏.基于改进的混合高斯模型的目标检测方法[J].计算机技术与发展,2012,(07):60.
　SHU Guang,ZHANG Chao,CAI Ze-su.Target Detection Method Based on Improved Gaussian Mixture Model[J].,2012,(07):60.
[4]刘洁,李目,周少武.一种混沌混合粒子群优化RBF神经网络算法[J].计算机技术与发展,2013,(08):181.
　LIU Jie[],LI Mu[],ZHOU Shao-wu[].An Algorithm of Chaotic Hybrid Particle Swarm Optimization Based on RBF Neural Network[J].,2013,(07):181.
[5]赵颖辉[][],蒋从锋[][]. 遥感影像的高性能并行处理技术研究[J].计算机技术与发展,2014,24(07):201.
　ZHAO Ying-hui[][],JIANG Cong-feng[][]. Research on High Performance Parallel Processing Technology for Remote Sensing Images[J].,2014,24(07):201.
[6]蒋翠清,孙富亮,吴艿芯. 基于相对欧氏距离的背景差值法视频目标检测[J].计算机技术与发展,2015,25(01):37.
　JIANG Cui-qing,SUN Fu-liang,WU Nai-xin. Video Object Detection of Background Subtraction Method Based on Relative Euclidean Distance[J].,2015,25(07):37.
[7]卢官明,衣美佳. 步态识别关键技术研究[J].计算机技术与发展,2015,25(07):100.
　LU Guan-ming,YI Mei-jia. Research on Critical Techniques in Gait Recognition[J].,2015,25(07):100.
[8]高翔,朱婷婷,刘洋. 多摄像头系统的目标检测与跟踪方法研究[J].计算机技术与发展,2015,25(07):221.
　GAO Xiang,ZHU Ting-ting,LIU Yang. Research of Target Detection and Tracking Method for Multi-camera System[J].,2015,25(07):221.
[9]章文洁[][],黄旻[],张桂峰[]. 滤光片多光谱成像中运动目标场景误配准修正[J].计算机技术与发展,2016,26(01):18.
　ZHANG Wen-jie[][],HUANG Min[],ZHANG Gui-feng[]. Misregistration Correction for Moving Object Scene in Filter-type Multispectral Imaging[J].,2016,26(07):18.
[10]张夏清,茅耀斌. 一种改进的ViBe背景提取算法[J].计算机技术与发展,2016,26(07):36.
　ZHANG Xia-qing,MAO Yao-bin. An Improved ViBe Background Generation Method[J].,2016,26(07):36.

更新日期/Last Update: 2022-07-10