窗口融合特征对比度的光学遥感目标检测

提出了一种基于窗口融合特征对比度的光学遥感目标检测方法。首先,在训练图像上生成大量不同尺寸的滑动窗, 计算了各窗口的多尺度显著度、仿射协变区域对比度、边缘密度对比度以及超像素完整度4项特征分值, 在确认集上基于窗口重合度和后验概率最大化学习各个特征的阈值参数。然后, 采用Naive Bayes框架进行特征融合, 并训练分类器。在目标检测阶段首先计算测试图像中各窗口的多尺度显著度分值, 初步筛选出显著度高且符合待检测目标尺寸比例的部分窗口。然后计算初选窗口集的其余3项特征, 再根据训练好的分类模型计算各个窗口的后验概率。最后, 挑选出局部高分值的候选区域并进行判断合并, 得到最终目标检测结果。针对飞机、油罐、舰船等3类遥感目标的检测结果显示: 4类特征在单独描述3类目标时表现出的性能各有差异, 最高检测准确率为74.21%~80.32%, 而融合方案能够综合考虑目标自身特点, 准确率提高至80.78~87.30%。与固定数量滑动窗方法相比, 准确率从约80%提高到约85%, 虚警率从20%左右降低为3%左右。最终高分值区域数降低约90%, 测试时间减少约25%。得到的结果显示该方法大大提高了目标检测精度和算法效率。

Abstract

A detection algorithm for optical remote sensing targets was proposed based on the fused features contrast of subwindows. Firstly, a large number of varisized sliding windows were generated in a training image, and four types of scores related to multi-scale saliency, affine invariant region contrast, edge density and superpixel straddling were computed within each window. The feature parameters were learned on validation sets by maximizing localization accuracy and posterior probability. Then, all the features were combined in a Naive Bayesian framework and a classifier was trained. In the target detection step, the multi-scale saliency score was firstly computed within all the windows of test images, and partial windows with higher saliency and proper sizes matching to the objects to be detected were selected preliminarily. Furthermore, other scores were computed within the selected windows, and the posterior probability of each window was computed by using the trained classifier. Finally, windows with high local scores were selected and merged and the final detection results were obtained. The detection experiments were performed on three types of remote targets including planes, oilcans and ships, and the results show that each type of feature appears different properties for targets described, the highest accuracy is 74.21% to 80.32%. The proposed method outperforms all the single feature methods and the accuracy is improved to 80.87% to 87.30%. By compared with the fixed number sliding window algorithm, the accuracy rate is improved from about 80% to 85% and the false alarm rate is reduced from about 20% to 3%. Furthermore, the proposed method shows a 90% reduction in the number of windows and 25% reduction in the detection time due to the selection in the intermediary stage. It concludes that the method improves detection accuracy and algorithm efficiency greatly.

参考文献

[1] DESAI C, RAMANNAN D, FOLKESS C. Discriminative models for multi-class object layout ［J］. International Journal of Computer Vision, 2011, 95(1): 1-12.

[2] FELZENSZWALB P F, GIRSHICK R B, DAVID M A, et al.. Object detection with discriminatively trained part-based models ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010,32(9): 1627-1645.

[3] LONG C, WANG X, HUA G, et al.. Accurate object detection with location relaxation and regionlets relocalization ［C］. Asian Conference on Computer Vision, Singapore, Nov 1-5, 2014: 260-275.

[4] LAMPERT C H, BLASCHKO M B, HOFMANN T. Beyond sliding windows: Object localization by efficient subwindow search ［C］. 2008 IEEE Conference on Computer Vision & Pattern Recognition, USA, June 24-26, 2008: 1-8.

[5] ZHANG S, XIE M. Beyond sliding windows: Object detection based on hierarchical segmentation model ［C］. IEEE International Conference on Communications,Hungary, Jun 9-13, 2013: 263-266.

[6] 张志龙, 杨卫平, 张淼, 等. 基于频谱残留变换的红外遥感图像舰船目标检测方法［J］. 电子与信息学报, 2015,37(9): 2144-2150.

ZHANG ZH L, YANG W P, ZHANG M, et al.. Ship detection in infrared remote sensing images based on spectral residual transform ［J］. Journal of Electronics & Information Technology, 2015, 37(9): 2144-2150. (in Chinese)

[7] 高立宁, 毕福昆, 龙腾, 等. 一种光学遥感图像海面舰船检测算法［J］. 清华大学学报(自然科学版), 2011, 51(1): 105-110.

GAO L N, BI F K, LONG T, et al.. Ship detection algorithm for optical remote sensing images ［J］. J Tsinghua Univ(Sci & Tech), 2011, 51(1): 105-110. (in Chinese)

[8] MOHAMMAD A, HOSSEIN E. Accurate object detection using local shape descriptors ［J］. Pattern Analysis and Application, 2015,18(2): 277-295.

[9] WANG X, LV Q, WANG B, et al.. Airport detection in remote sensing images: a method based on saliency map ［J］. Cognitive Neurodynamics, 2013, 7(2): 143-154.

[10] 聂海涛,龙科慧,马军,等.采用改进尺度不变特征变换在多变背景下实现快速目标识别［J］.光学精密工程, 2015, 23(8): 2349-2355.

NIE H T, LONG K H, MA J, et al.. Fast object recognition under multiple varying background using improved SIFT method ［J］. Opt. Precision Eng., 2015, 23(8): 2349-2355. (in Chinese)

[11] 周姣, 辛云宏. 基于显著性与尺度空间的红外弱小目标检测［J］. 激光与红外, 2015, 45(4): 452-456.

ZHOU J, XIN Y H. Infrared dim small target detection based on saliency and scale-space ［J］. Laser & Infrared, 2015, 45(4): 452-456. (in Chinese)

[12] 张宇, 何楚, 石博, 等. 逐层特征选择的多层部件模型用于遥感图像飞机目标检测［J］. 武汉大学学报(信息科学版), 2014, 39(12): 1406-1411.

ZHANG Y, HE CH, SHI B, et al.. Multi-layer feature selection based hierarchal component model for aero-plane detection on remote sensing image ［J］. Geomatics and Information Science of Wuhan University, 2014, 39(12): 1406-1411. (in Chinese)

[13] 安彧, 王小非, 夏学知, 等. 海战场光学遥感图像舰船目标检测［J］. 武汉大学学报(工学版), 2015, 48(4): 568-573.

AN H, WANG X F, XIA X ZH, et al.. Detection of sea battlefield’s ship targets in optical remote sensing images ［J］. Engineering Journal of Wuhan University, 2015, 48(4): 568-573. (in Chinese)

[14] LI Y, WANG S, TIAN Q, et al.. Learning cascaded shared-boost classifiers for part-based object detection ［J］. IEEE Transactions on Image Processing, 2014, 23(4): 1858-1871.

[15] LEIBE B, LEONARDIS A, SCHIELE B. Robust object detection with interleaved categorization and segmentation ［J］. International Journal of Computer Vision,2015,77(1-3): 259-289.

[16] 李平,魏仲慧,何昕,等. 采用多形状特征融合的多视点目标识别［J］. 光学精密工程,2014, 22(12): 3368-3376.

LI P, WEI ZH H,HE X, et al.. Object recognition based on shape fearure fusion under multi-views ［J］. Opt. Precision Eng., 2014, 22(12): 3368-3376. (in Chinese)

[17] 谭熊,余旭初,张鹏强,等.基于多核支持向量机的高光谱影像非线性混合像元分解［J］.光学精密工程, 2014, 22(7): 1912-1920.

TAN X, YU X CH, ZHANG P Q, et al.. NonLinear mixed pixel decomposition of hyperspectral imagery based on multiple kernel SVM ［J］. Opt. Precision Eng., 2014, 22(7): 1912-1920. (in Chinese)

[18] 杜杰, 吴谨, 朱磊. 基于区域特征融合的RGBD显著目标检测［J］.液晶与显示, 2016, 31(1): 117-123.

DU J, WU J, ZHU L, RGBD salient object detection based on regional feature integration［J］. Chinese Journal of Liquid Crystals and Displays, 2016, 31(1): 117-123. (in Chinese)

[19] 邓丹, 吴谨, 朱磊, 等. 基于纹理抑制和连续分布估计的显著性目标检测方法［J］.液晶与显示, 2015, 30(1): 120-125.

DENG D, WU J, ZHU L, et al.. Significant target detection method based on texture inhibition and continuous distribution estimation ［J］. Chinese Journal of Liquid Crystals and Displays, 2015, 30(1): 120-125. (in Chinese)

[20] HOU X, ZHANG L. Saliency detection: A spectral residual approach［C］. IEEE Conference on Computer Vision and Pattern Recognition, 2007: 1-8.

[21] MIKOLAJCZYK K, SEHMID C. Scale & affine invariant interest point detectors ［J］. International Journal on Computer Vision, 2004, 60(1): 63-86.

[22] ALEXE B, DESELAERS T, FERRARI V. Measuring the objectness of image windows ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence,2012, 34(11): 2189-2202.

[23] MARTIN D, FOWLKES C, MALIK J. Learning to detect natual image boundaries using local brightness, color, and textures cues［J］. PAMI, 2004, 26(5): 530-549.

[24] FELZENSZWALB P F, HUTTENLOCHER D P. Efficient graph-based image segmentation［J］. International Journal on Computer Vision, 2004, 59(2): 167-181.

[25] 孙显, 王宏琦, 杨志峰. 基于形状统计模型的多类目标自动识别方法［J］. 电子与信息学报, 2009, 31(11): 2626-2631.

SUN X, WANG H Q, YANG ZH F. Automatic multi-categorical objects recognition using shape statistical models［J］. Journal of Electronics & Information Technology, 2009, 31(11): 2626-2631. (in Chinese)

李湘眷, 王彩玲, 李宇, 孙皓. 窗口融合特征对比度的光学遥感目标检测[J]. 光学精密工程, 2016, 24(8): 2067. LI Xiang-juan, WANG Cai-ling, LI Yu, SUN Hao. Optical remote sensing object detection based on fused feature contrast of subwindows[J]. Optics and Precision Engineering, 2016, 24(8): 2067.

窗口融合特征对比度的光学遥感目标检测

关于本站 Cookie 的使用提示

全站搜索

窗口融合特征对比度的光学遥感目标检测

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索