一种改进的LBP特征实现铁路扣件识别

王强; 李柏林; 侯云; 范宏

doi:10.3969/j.issn.0258-2724.2018.05.003

一种改进的LBP特征实现铁路扣件识别

doi: 10.3969/j.issn.0258-2724.2018.05.003

王强^1,2,,
李柏林^1, ,,
侯云¹,
范宏¹

详细信息

作者简介:
王强（1979—），男，博士研究生，研究方向为计算机视觉、图像处理，E-mail: wqiangsky@163.com

通讯作者:
李柏林（1962—），男，教授，博士生导师，研究方向为计算机图形图像处理，E-mail: blli62@263.net

中图分类号: TP391.41
计量
- 文章访问数: 558
- HTML全文浏览量: 264
- PDF下载量: 77
- 被引次数: 0
出版历程
- 收稿日期: 2017-08-08
- 刊出日期: 2018-10-01

An Improved LBP Feature for Rail Fastener Identification

WANG Qiang^{1,2
,},
LI Bailin^{1
, ,},
HOU Yun¹,
FAN Hong¹

摘要

摘要: 为了提高铁路扣件基于视觉的自动化检测精度，提出了一种改进的LBP （local binary pattern）编码算法. 该方法为了避免基本LBP对噪声敏感问题，根据不同邻域的不同噪声强度，结合测量误差服从高斯分布的原则，计算邻域内像素均值和偏差；根据偏差大小，自动设置阈值，实现自适应噪声抑制. 为了避免基本LBP表达邻域差分关系不完整的缺陷，提出了利用邻域内随机采样的方式得到采样点对，通过比较随机点对的差分关系得到LBP编码. 对在晴天、阴天、雨天等不同天气条件下的铁路扣件图像上进行实验，并与原始以及其他改进LBP进行比较. 结果表明，本文的算法具有更高的检测准确率，晴天提高了3.32%，阴天提高了3.27%，雨天提高了4.10%，能够满足铁路扣件自动化检测的需要.
- LBP /
- 抗噪声 /
- 扣件 /
- 图像识别 /
- 图像检测
Abstract: An improved LBP (local binary pattern) algorithm is proposed to raise the auto-detection accuracy on fasteners. The original LBP is sensitive to noise. To solve this problem, the pixel means value and deviation are calculated, according to the different noise in different neighborhood and the measuring error is following the Gaussian distribution. Then the threshold is set to realize adaptive noise suppression, according to the deviation values. The original LBP cannot completely express the neighboring difference relationship. To avoid this defect, the proposed method gets the sampled point pairs by random sampling in neighborhood. Then LBP coding is generated through comparing the difference relationship of random point pairs. Tests are carried on fasteners images on clear, cloudy and rainy days with original and other improved LBP algorithms. The comparing results show the proposed method owns more detection accuracy. The detection accuracies increase by 3.32%, 3.27% and 4.10% independently on clear, cloudy and rainy days, which shows the proposed method can meet auto-detection on railway fasteners.
- LBP /
- anti-noise /
- fastener /
- image recognition /
- image detection

HTML全文

图 1 噪声抑制过程示意

Figure 1. Schematic diagram of noise suppression process

下载: 全尺寸图片幻灯片

图 2 随机选择点对LBP编码示意

Figure 2. Schematic diagram of LBP coding for random point pairs

下载: 全尺寸图片幻灯片

图 3 随机点生成示意

Figure 3. Schematic diagram for random points generating

下载: 全尺寸图片幻灯片

图 4 不同天气下扣件

Figure 4. Fastener images under different weathers

下载: 全尺寸图片幻灯片

图 5 本文编码图像示例

Figure 5. Schematic diagram for coding images

下载: 全尺寸图片幻灯片

图 6 LBP编码图像对比示例

Figure 6. Contrast examples for coding images

下载: 全尺寸图片幻灯片

图 7 扣件图像不同阈值下3种编码方式的识别率

Figure 7. Fastener Recognition rates of three coding methods with different thresholds

下载: 全尺寸图片幻灯片

表 1 晴天扣件识别率

Table 1. Fastener recognition on clear day

方法	识别率/%
LBP	89.35
LBP^U2	90.46
LTP	91.85
DLBP	90.89
FLBP	91.42
NRLBP	93.62
本文方法	96.94

下载: 导出CSV

表 3 雨天扣件识别率

Table 3. Fastener recognition on rainy day

方法	识别率/%
LBP	75.16
LBP^U2	77.32
LTP	80.52
DLBP	79.06
FLBP	79.45
NRLBP	81.23
本文方法	85.33

下载: 导出CSV

表 2 阴天扣件识别率

Table 2. Fastener recognition on cloudy day

方法	识别率/%
LBP	85.76
LBP^U2	86.59
LTP	88.03
DLBP	86.95
FLBP	88.72
NRLBP	89.57
本文方法	92.84

下载: 导出CSV

表 4 计算时间对比

Table 4. Computation time comparisons

方法	时间/s
LBP	108.711
本文方法	157.408

下载: 导出CSV

参考文献(22)

SINGH M, SINGH S, JAISWAL J, et al. Autonomous rail track inspection using vision based system[C]//IEEE International Conference on Computational Intelligence for Homeland Security and Personal Safety. [S. l.]: Alexandria: IEEE, 2006: 56-59

MARINO F, DISTANTE A, MAZZEO P, et al. A real-time visual inspection system for railway maintenance: automatic hexagonal-headed bolts detection[J]. IEEE Transactions on Systems,Man,and Cybernetics,Part C (Applications and Reviews), 2007, 37(3): 418-428

YELLA S, DOUGHERTY M, GUPTA N. et al. Condition monitoring of wooden railway sleepers[J]. Transportation Research Part C: Emerging Technologies, 2009, 17(1): 38-55

许贵阳,史天运,任盛伟,等. 基于计算机视觉的车载轨道巡检系统研制[J]. 中国铁道科学,2013,34(1): 139-144

XU Guiyang, SHI Tianyun, REN Shengwei, et al. Development of the on-board track inspection system based on computer vision[J]. China Railway Science, 2013, 34(1): 139-144

肖新标,金学松,温泽峰. 钢轨扣件失效对列车动态脱轨的影响[J]. 交通运输工程学报,2006,6(1): 10-15

XIAO Xinbiao, JIN Xuesong, WEN Zefeng. Influence of rail fastener failure on vehicle dynamic derail-ment[J]. Journal of Traffic and Transportation Engineering, 2006, 6(1): 10-15

YANG Jinfeng, WEI Tao, LIU Manhua, et al. An efficient direction field-based method for the detection of fasteners on high-speed railways[J]. Sensors, 2011, 11(8): 7364-7381

XIA Yiqi, XIE Fengying, JIANG Zhiguo. Broken railway fastener detection based on adaboost algorithm[C]// International Conference on Opto-electronics and Image Processing. Beijing: IEEE, 2010: 313-316

刘甲甲,李柏林,罗建桥,等. 融合PHOG和MSLBP特征的铁路扣件检测算法[J]. 西南交通大学学报,2015,50(2): 256-263

LIU Jiajia, LI Bailin, LUO Jianqiao. et al. Railway fastener detection algorithm integrating PHOG and MSLBP features[J]. Journal of Southwest Jiaotong University, 2015, 50(2): 256-263

REN J, JIANG X, YUAN J. LBP encoding schemes jointly utilizing the information of current bit and other LBP bits[J]. IEEE Signal Processing Letters, 2015, 22(12): 2373-2377

REN J, JIANG X, YUAN J. Noise-resistant local binary pattern with an embedded error-correction mechanism[J]. IEEE Transactions on Image Processing, 2013, 22(10): 4049-4060

GUO Zhenhua, ZHANG Lei, ZHANG D. A completed modeling of local binary pattern operator for texture classification[J]. IEEE Transactions on Image Processing, 2010, 19(6): 1657-1663

SONG Tiecheng, LI Hongliang, MENG Fanman, et al. Noise-robust texture description using local contrast patterns via global measures[J]. IEEE Signal Processing Letters, 2014, 21(1): 93-96

WANG Kai, BICHOT C, ZHU Chao, et al. Pixel to patch sampling structure and local neighboring intensity relationship patterns for texture classification[J]. IEEE Signal Processing Letters, 2013, 20(9): 853-856

LIAO S, LAW M W K, CHUNG A C S. Dominant local binary patterns for texture classification[J]. IEEE transactions on image processing, 2009, 18(5): 1107-1118

KWAK J T, XU S, WOOD B J. Efficient data mining for local binary pattern in texture image analysis[J]. Expert Systems with Applications, 2015, 42(9): 4529-4539

SATPATHY A, JIANG X, ENG H L. LBP-based edge-texture features for object recognition[J]. IEEE Transactions on Image Processing, 2014, 23(5): 1953-1964

REN J, JIANG X, YUAN J. et al. Optimizing LBP structure for visual recognition using binary quadratic programming[J]. IEEE Signal Processing Letters, 2014, 21(11): 1346-1350

LI Wei, CHEN Chen, SU Hongjun, et al. Local binary patterns and extreme learning machine for hyperspectral imagery classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(7): 3681-3693

OJALA T, PIETIKAINEN M, MAENPAA T. Multiresolution gray-scale and rotation invariant texture classification with local binary patterns[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(7): 971-987

TAN X, TRIGGS B. Enhanced local texture feature sets for face recognition under difficult lighting conditions[J]. IEEE Transactions on Image Processing, 2010, 19(6): 1635-1650

AHONEN T, PIETIKAINEN M. Soft histograms for local binary patterns[C]//Proceedings of the Finnish signal processing symposium. Finsic: [s. n.], 2007, 5(9): 1-4

WEINBERGER M J, RISSANEN J J, AERS R B. Applications of universal context modeling to lossless compression of gray-scale images[J]. IEEE Transactions on Image Processing, 1996, 5(4): 575-586

施引文献

附加材料(0)

访问统计