文档介绍：复杂背景噪声条件下图像识别技术初探
摘要
随着信息技术和数字图像处理技术的发展,交通管理系统的信息化、数字化是大势所趋。路面裂纹识别系统(PCR)是数字化交通管理系统(DTS)的重要组成部分,在现代公路检测管理系统中占有举足轻重的作用。
近年来,使用计算机来对图像进行处理和分析,已经获得了飞速的发展。对路面裂纹识别系统中关键技术的研究已经成为科学界的一个热点问题。
本文研究了一种通过构造8 个方向模板的Sobel 算子对路面裂纹进行检测与识别的算法。由于经典图像处理算法对路面裂纹的检测效果不佳, 以及神经网络等智能算法不易实现的弊端, 提出了一种新的路面裂纹检测的图像处理算法,该算法结合迭代阈值分割算法和全方位膨胀形态学方法对边缘检测后的图像进行处理, 能够很好地获得路面的裂纹特性。最后, 用MATLAB 图像处理工具进行算法实现。结果表明, 裂缝边缘的连接和检测效果较好, 为路面裂纹的分析和管理提供了技术基础。
关键词:图像处理;路面裂纹检测; 形态学; Sobel 算子
RESEARCH ON IMAGE RECOGNITION PLEX NOISE BACKGROUND
ABSTRACT
With the developing of information technology and digital image processing technology, informazation and digitalization of traffic management is the trend. Pavement Cracks Recognition(PCR) system is the core of Digital Traffic Systems(DTS). It is very important in modern traffic management systems.
In recent years, use puter to deal with and analyses image has already obtained progress at full speed. The study on the critical techniques for PCR has already e an important research field of scientific circles.
It is bad to detect pavement surface cracks by classical image processing algorithms and very difficult to realize the image processing algorithms by intelligent algorithms, such as work. In order to solve these questions, a new image processing algorithm is presented by analyzing characteristics of the pavement surface cracks in this paper. Our algorithm constructs Sobel operator of 8 orientations’ template to detect the pavement surface cracks, and employs the segmentation algorithm of the iterative threshold and the expanded morphological algorithm of all orientations to process the detected images. By making use of the above algorithm, we can well obtain the characteristics of the pavement surface cracks. At last, we use the tools of image processing of MATLAB to realize the algorithm. The results show that the juncture and detection of the crack edges are effectively realized. And provide the techno-guarantees for the analysis and management of