摘要
提出了一种将组合特征提取和SVM相结合的山体滑坡识别算法。首先进行图像大小统一和对比度增强的预处理;然后在纵向划分子区域的基础上,采用小波尺度共生矩阵对各子区域进行纹理特征提取、在HSI颜色空间中对各子区域进行颜色特征提取、在RGB颜色空间中进行分割特征提取;最后将组合特征送入分类器进行分类,对是否发生滑坡进行识别。实验表明,该方法研究解决的铁路线山体滑坡灾害图像识别技术,为实现铁路沿线环境自动识别奠定了基础。
To realize the recognition of landslide disaster,this paper proposes a algorithm of landslides recognition based on combined features and SVM.Firstly,the image shall be preprocessed,including size equalization and histogram equalization.And then divide the image into sub-regions vertically,extracted the texture features of Scale Co-occurrence Matrix based on Wavelet transform in each sub-region,extract the color features based on HSI color space in each sub- region,and extract the image segmentation feature based on RGB color space.At last,sent the combined features into classifier to classify and recognize.Experiments show that the image recognition technology of landslide disaster along railway provides foundation for automatic recognition of surroundings along railway.
出处
《电子测量技术》
2013年第8期65-70,共6页
Electronic Measurement Technology
