水利水电技术2017，Vol.48Issue(1)：40-45,6.DOI:10.13928/j.cnki.wrahe.2017.01.008
基于最优特征选择与支持向量机的钱塘江涌潮检测算法
Optimal feature selection and support vector machines-based algorithm for detection of tidal bore in Qiantangjing River
摘要
Abstract
By taking the design of a fully new background model algorithm as the objective,a feature template is composed by the suitable pixel features selected from the relevant feature pool with the probability densities obtained by Gaussian kernel function through counting the density estimated values among various intervals,and then both the foreground and the background can be separated through the comparison with a fixed number of the pixel eigenvalues trained by the support vector machines by taking the pixel eigenvalues at the corresponding positions in the input video-streams as the input.The result from the application of this method to the detection of the tidal bore in Qiangtangjiang River shows that all the F-measure values are over 65％ with stronger robustness and the recognition rate for selecting radial basis function with support vector machine is over 90％ with higher computing speed,and then it can reduce the disturbance from the fluctuation of water surface with a high accuracy,thus can provide an important tool for description of the dynamic characteristics of the river.关键词
最优特征选择/支持向量机/背景建模/运动目标检测/涌潮检测/钱塘江Key words
optimal feature selection/support vector machines/background modeling/moving object detection/tidal bore detection/Qiangtangjiang River分类
计算机与自动化引用本文复制引用
高鹏,王瑞荣,王培力..基于最优特征选择与支持向量机的钱塘江涌潮检测算法[J].水利水电技术,2017,48(1):40-45,6.基金项目
国家自然科学基金项目(61374005) （61374005）
浙江省自然科学基金项目(LY14F030022) （LY14F030022）
