摘要
铁路扣件是固定轨道的连接件,扣件丢失或断裂直接影响铁路运输的安全。扣件运行环境复杂,采集的图像与背景差异性较小,难以自动识别。针对扣件图像自动化识别提出了一种新的图像局部二元模式编码算法。用高斯分布进行随机采样得到邻域点,利用随机点对的差分关系得到二元模式编码,称为高斯采样二元模式编码GSLBP (Gaussian sampling local binary pattern)。为了避免噪声影响,利用随机点邻域像素值之和代替随机点的值,通过比较像素值之和得到编码,最后利用卡方距离对图像进行分类。该算法更加准确的反映了图像局部纹理信息,图像差异化信息更加明显。在铁路扣件图像上进行了实验,比较了各种LBP编码方法与提出的方法的分类结果,结果表明提出的方法具有更好的分类结果。
Railway fasteners are the connectors of the fixed railway track. Fasteners lost or broken directly affect the safety of railway transport. The complicated running environment,the indistinguishable differences between fasteners images and background make it difficult to automatically identify the fasteners. A new local binary image coding algorithm is proposed to automatically identify the fasteners. Gaussian sampling was used to obtain the neighborhood point randomly,and difference relation of random point pairs was adopted to get the binary pattern encoding,which was called Gaussian sampling local binary pattern. In order to avoid the influence of noise,the sum of the pixel values of the random point was substituted for the value of the random point,and the sum of the pixel values was compared to get the encoding. Finally,the image was classified with chi-square distance. The algorithm reflects the image of local texture information more accurately and image differentiation information is more obvious. Tests on railway fasteners of the LBP coding method and the proposed method were carried and compared. The results show that the proposed method has better classification results.
作者
王强
李柏林
侯云
WANG Qiang;LI Bai-lin;HOU Yun(SOUTHwest Jiaotong University,School of Mechanical Engineering,Chengdu Sichuan 610031,China;Chengdu Technological University,College of Mechanieal Engineering,Chengdu Sichuan 611730,China)
出处
《计算机仿真》
北大核心
2018年第11期421-425,435,共6页
Computer Simulation
基金
国家自然科学基金(51275431)
四川省科技厅项目(2016GZ0194)
四川省教育厅项目(16ZB0330)
四川省大学生创新创业项目(201611116007)
关键词
局部二元模式
图像识别
图像分类
高斯分布
随机采样
LBP
Image recognition
Image classification
Gaussian distribution
Random sampling
