摘要
针对大多数路面裂缝检测算法对坑洼、松散和车辙等复杂病害分割适应较差的问题,提出了一种基于训练样本自动选取与改进的核极限学习机相结合的检测方法。首先使用二维Otsu选取训练样本并提取LBP特征和HOG特征,然后采用遗传算法对核极限学习机中随机给定的输入权值和隐含层偏差进行优化,将降维后得到的特征向量作为特征属性对改进的核极限学习机进行训练,最后用训练好的分类器对路面病害进行检测。经实验证明,该算法与对比实验相比分割精度提高了24.8%,运行时间为4.31 s,是一种鲁棒性较强的检测方法。
Most of the pavement crack detection algorithms have poor segmentation effect on complex diseases such as potholes, looseness and ruts. A detection method based on automatic selection of training samples and improved kernel extreme learning machine is proposed. Firstly, two-dimensional Otsu is used to select training samples and extract LBP features and HOG features. Then, genetic algorithm is used to optimize the input weight and hidden layer threshold given randomly in kernel extreme learning machine. The feature vectors obtained after dimension reduction are used as feature attributes to train the improved kernel extreme learning machine. Finally, the trained classifier is used to detect the pavement diseases. The experimental results show that the segmentation accuracy of the algorithm is improved by 24.8% and the running time is 4.31 s, which is a robust detection method.
作者
李鹏
王青宁
单钰强
LI Peng;WANG Qingning;SHAN Yuqiang(Jiangsu Collaborative Innovation Center qf Atmospheric Environment and Equipment Technology,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China;Jiangsu Key Laboratory of Meteorological Observation and Information Processing,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China;Binjiang College,Nanjing University of Information Science and Technology,Wuxi Jiangsu 214105,China)
出处
《电子器件》
CAS
北大核心
2022年第1期143-147,共5页
Chinese Journal of Electron Devices
基金
国家自然科学基金项目(41075115)
江苏省重点研发计划社会发展项目(BE2015692)
江苏省第11批六大高峰人才项目(2014-XXRJ-006)。
关键词
特征提取
主成分分析
核极限学习机
feature extraction
principal component analysis
kernel extreme learning machine
