一级黄色片免费播放|中国黄色视频播放片|日本三级a|可以直接考播黄片影视免费一级毛片

高級搜索

留言板

尊敬的讀者、作者、審稿人, 關(guān)于本刊的投稿、審稿、編輯和出版的任何問題, 您可以本頁添加留言。我們將盡快給您答復(fù)。謝謝您的支持!

姓名
郵箱
手機號碼
標(biāo)題
留言內(nèi)容
驗證碼

基于生成對抗網(wǎng)絡(luò)和線上難例挖掘的SAR圖像艦船目標(biāo)檢測

李健偉,  曲長文,  彭書娟,  江源

李健偉, 曲長文, 彭書娟, 江源. 基于生成對抗網(wǎng)絡(luò)和線上難例挖掘的SAR圖像艦船目標(biāo)檢測[J]. 電子與信息學(xué)報, 2019, 41(1): 143-149. doi: 10.11999/JEIT180050
引用本文: 李健偉, 曲長文, 彭書娟, 江源. 基于生成對抗網(wǎng)絡(luò)和線上難例挖掘的SAR圖像艦船目標(biāo)檢測[J]. 電子與信息學(xué)報, 2019, 41(1): 143-149. doi: 10.11999/JEIT180050
Jianwei LI, Changwen QU, Shujuan PENG, Yuan JIANG. Ship Detection in SAR images Based on Generative Adversarial Network and Online Hard Examples Mining[J]. Journal of Electronics & Information Technology, 2019, 41(1): 143-149. doi: 10.11999/JEIT180050
Citation: Jianwei LI, Changwen QU, Shujuan PENG, Yuan JIANG. Ship Detection in SAR images Based on Generative Adversarial Network and Online Hard Examples Mining[J]. Journal of Electronics & Information Technology, 2019, 41(1): 143-149. doi: 10.11999/JEIT180050

基于生成對抗網(wǎng)絡(luò)和線上難例挖掘的SAR圖像艦船目標(biāo)檢測

doi: 10.11999/JEIT180050 cstr: 32379.14.JEIT180050
  • 1.

    海軍航空大學(xué) ??煙臺 ??264001

  • 2.

    92228部隊 ??北京 ??100044

基金項目: 國家自然科學(xué)基金(61571454)
詳細信息
    作者簡介:

    李健偉：男，1989年生，博士生，研究方向為SAR圖像處理、機器學(xué)習(xí)及深度學(xué)習(xí)

    曲長文：男，1964年生，教授，研究方向為雷達信號處理，信息對抗，信號與信息處理等

    彭書娟：女，1980年生，博士生，研究方向為SAR圖像處理

    通訊作者:

    李健偉　lgm_jw@163.com

  • 中圖分類號: TN957.51

Ship Detection in SAR images Based on Generative Adversarial Network and Online Hard Examples Mining

  • 1.

    Naval Aeronautical University, Yantai 264001, China

  • 2.

    The Unit 92228 of PLA, Beijing 100044, China

Funds: The National Natural Science Foundation of China (61571454)
  • 摘要:

    基于深度學(xué)習(xí)的SAR圖像艦船目標(biāo)檢測算法對圖像的數(shù)量和質(zhì)量有很高的要求，而收集大體量的艦船SAR圖像并制作相應(yīng)的標(biāo)簽需要消耗大量的人力物力和財力。該文在現(xiàn)有SAR圖像艦船目標(biāo)檢測數(shù)據(jù)集(SSDD)的基礎(chǔ)上，針對目前檢測算法對數(shù)據(jù)集利用不充分的問題，提出基于生成對抗網(wǎng)絡(luò)(GAN)和線上難例挖掘(OHEM)的SAR圖像艦船目標(biāo)檢測方法。利用空間變換網(wǎng)絡(luò)在特征圖上進行變換，生成不同尺寸和旋轉(zhuǎn)角度的艦船樣本的特征圖，從而提高檢測器對不同尺寸、旋轉(zhuǎn)角度的艦船目標(biāo)的適應(yīng)性。利用OHEM在后向傳播過程中發(fā)掘并充分利用難例樣本，去掉檢測算法中對樣本正負比例的限制，提高對樣本的利用率。通過在SSDD數(shù)據(jù)集上的實驗證明以上兩點改進對檢測算法性能分別提升了1.3%和1.0%，二者結(jié)合提高了2.1%。以上兩種方法不依賴于具體的檢測算法，且只在訓(xùn)練時增加步驟，在測試時候不增加計算量，具有很強的通用性和實用性。

    Abstract:

    Deep learning based ship detection method has a strict demand for the quantity and quality of the SAR image. It takes a lot of manpower and financial resources to collect the large volume of the image and make the corresponding label. In this paper, based on the existing SAR Ship Detection Dataset (SSDD), the problem of insufficient utilization of the dataset is solved. The algorithm is based on Generative Adversarial Network (GAN) and Online Hard Examples Mining (OHEM). The spatial transformation network is used to transform the feature map to generate the feature map of the ship samples with different sizes and rotation angles. This can improve the adaptability of the detector. OHEM is used to discover and make full use of the difficult sample in the process of backward propagation. The limit of positive and negative proportion of sample in the detection algorithm is removed, and the utilization ratio of the sample is improved. Experiments on the SSDD dataset prove that the above two improvements improve the performance of the detection algorithm by 1.3% and 1.0% respectively, and the combination of the two increases by 2.1%. The above two methods do not rely on the specific detection algorithm, only increase the time in training, and do not increase the amount of calculation in the test. It has very strong generality and practicability.

    • HTML全文

  • 圖  1  Fast R-CNN原理示意圖

    圖  2  對抗空間變換網(wǎng)絡(luò)示意圖

    圖  3  線上難樣本挖掘流程(K=64)

    圖  4  數(shù)據(jù)集SSDD中不同尺度和角度艦船檢測結(jié)果

    表  1  4種方法檢測性能

    方法mAP
    (%)    		訓(xùn)練時間
    (s)    		測試時間
    (s)
    標(biāo)準(zhǔn)的 Fast R-CNN68.00.6100.328
    標(biāo)準(zhǔn)的 Fast R-CNN+ GAN69.40.8230.326
    標(biāo)準(zhǔn)的 Fast R-CNN+OHEM69.11.1520.321
    標(biāo)準(zhǔn)的 Fast R-CNN+GAN
    +OHEM    		70.22.1090.330
    下載: 導(dǎo)出CSV
  • KRIZHEVSKY A, SUTSKEVER I and HINTON G E. ImageNet classification with deep convolutional neural networks[C]. International Conference on Neural Information Processing Systems. Nevada, USA, 2012: 1097–1105.
    GIRSHICK, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]. IEEE Conference on Computer Vision and Pattern Recognition, Columbus, USA, 2014: 580–587.
    REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: Towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 39(6): 1137–1149. doi: 10.1109/TPAMI.2016.2577031
    REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: Unified, real-time object detection[C]. IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 779–788.
    LIU Wei, ANGUELOV D, ERHAN D, et al. SSD: Single shot multiBox detector[C]. IEEE European Conference on Computer Vision, Amsterdam, Netherlands, 2016: 21–37.
    王思雨, 高鑫, 孫皓, 等. 基于卷積神經(jīng)網(wǎng)絡(luò)的高分辨率SAR圖像飛機目標(biāo)檢測方法[J]. 雷達學(xué)報, 2017, 6(2): 195–203. doi: 10.12000/JR17009

    Wang Siyu, Gao Xin, Sun Hao, et al. An aircraft detection method based on convolutional neural networks in high-resolution SAR images[J]. Journal of Radars, 2017, 6(2): 195–203. doi: 10.12000/JR17009
    徐豐, 王海鵬, 金亞秋. 深度學(xué)習(xí)在SAR目標(biāo)識別與地物分類中的應(yīng)用[J]. 雷達學(xué)報, 2017, 6(2): 136–148. doi: 10.12000/JR16130

    Xu Feng, Wang Haipeng, and Jin Yaqiu. Deep learning as applied in SAR target recognition and terrain classification[J]. Journal of Radars, 2017, 6(2): 136–148. doi: 10.12000/JR16130
    劉澤宇, 柳彬, 郭煒煒, 等. 高分三號NSC模式SAR圖像艦船目標(biāo)檢測初探[J]. 雷達學(xué)報, 2017, 6(5): 473–482. doi: 10.12000/JR17059

    Liu Zeyu, Liu Bin, Guo Weiwei et al. Ship detection in GF-3 NSC mode SAR images[J]. Journal of Radars, 2017, 6(5): 473–482. doi: 10.12000/JR17059
    HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C]. IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 770–778.
    HUANG Gao, LIU Zhuang, WEINBERGER K Q, et al. Densely connected convolutional networks[C]. IEEE Conference on Computer Vision and Pattern Recognition, Hawaii, USA, 2017: 4700–4708.
    SUNG K K and POGGIO T. Example-based learning for view-based human face detection[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2002, 20(1): 39–51. doi: 10.1109/34.655648
    SHRIVASTAVA A, GUPTA A, and GIRSHICK R. Training region-based object detectors with online hard example mining[C]. IEEE Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 761–769.
    UIJLINGS J R R, SANDE K, GEVES T, et al. Selective search for object recognition[J]. International Journal of Computer Vision, 2013, 104(2): 154–171. doi: 10.1007/s11263-013-0620-5
    WANG Xiaolong, SHRIVASTAVA A, and GUPTA A. A-Fast-RCNN: Hard positive generation via adversary for object detection[C]. IEEE Conference on Computer Vision and Pattern Recognition, Hawaii, USA, 2017.
    JADERBERG M, KAREN S, and ANDREW Z. Spatial transformer networks[C]. Advances in Neural Information Processing Systems, Montreal, Canada, 2015: 2017–2025.
    GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets[C]. International Conference on Neural Information Processing Systems, Montreal, Canada, 2014: 2672–2680.
  • 加載中
圖(4) / 表(1)
計量
  • 文章訪問數(shù):  2592
  • HTML全文瀏覽量:  1084
  • PDF下載量:  183
  • 被引次數(shù): 0
出版歷程
  • 收稿日期:  2018-01-15
  • 修回日期:  2018-09-26
  • 網(wǎng)絡(luò)出版日期:  2018-10-22
  • 刊出日期:  2019-01-01

目錄

    /

    返回文章
    返回