| 摘 要: 将计算机机器视觉、图像处理及图像分析技术紧密结合起来并用于图像纹理特征的分析，通过总结已有的研究成果，研究分析图像在区域化、边缘化、数学变换化等不同算法思想下提取的纹理特征，并总结提出了一种计算成本趋向实时处理的图像纹理特征处理方案。该方案研究了更为精细的纹理分类、纹理分割、纹理拼接和纹理配准等算法，并可实现图像纹理特征提取的准确性和良好的扩展性，实现后续分类、分割、拼接和配准等基本应用的及时性、准确性和高效性。 |
| 关键词: 图像纹理特征分析；计算机机器视觉；图像处理 |
|
中图分类号: TP391.4
文献标识码: A
|
|
| Image Texture Feature Analysis and Extraction Method |
|
LEI Yuguo1,2, LIANG Nan1,2, LIU Chunmei1,2
|
( 1.Institute of Applied Physics, Henan Academy of Sciences, Zhengzhou 450008, China ;
2.Henan Province Internet of Things Perception Technology and Systems Key Laboratory, Zhengzhou 450008, China)
710952005@qq.com; 149016898@qq.com; 49072544@qq.com
|
| Abstract: This paper proposes to closely combine computer machine vision, image processing and image analysis technologies, which is used for analyzing image texture features. By summarizing the existing research results, image texture features, which are extracted by different algorithm ideas such as regionalization, marginalization and mathematical transformation, are studied and analyzed. An image texture feature processing scheme whose computational cost tends to be processed in real time is proposed. More refined algorithms of texture classification, texture segmentation, texture splicing and texture registration are studied in this scheme, which can achieve the accuracy and good scalability of image texture feature extraction. This scheme also achieve timeliness, accuracy and efficiency of subsequent essential applications, such as classification, segmentation, splicing and registration, etc.. |
| Keywords: analysis of image texture features; computer machine vision; image processing |
