信息工程毕业论文 基于数字图像的直线检测技术研究.docVIP

基于数字图像的直线检测技术研究 摘 要 直线检测是图像处理中十分重要的内容之一，是图像分割的基础，在道路识别，建筑物识别，医学图像分析等领域都有十分重要的应用。在图像处理领域，直线特征经常被用作高层处理，如立体匹配、目标跟踪、物体识别等，所以直线检测和分析对于数字图像处理有着重要的意义。 直线,特别是线段,有简单的几何特性,用它们描述目标的边缘特征可以方便对规则的物体进行分类,进而可以提高后续图像处理的效率。尽管有相当多的研究者投入了大量的精力在直线检测方面做了很多工作，也研究出了不少算法，如链码检测，DOG函数，最小二乘法，radon变换等，但是目前直线检测最成熟、应用最广泛的算法仍然是Hough变换。Hough变换是Paul Hough在1962年提出的,此后很多人对经典Hough变换算法进行了改进,使得Hough变换的得到了更好的运用。而利用Hough变换鲁棒性好，提取稳定，精度高，抗噪性强等优点来检测图像的直线特征,不但可以识别简单背景中的规则几何图形,还可以检测复杂背景中真实目标的直线特征。所以本课题就采用Hough变换对图像进行直线检测。 关键词：Hough变换 直线检测 图像处理 Research based on digital image of line detection technology ABSTRACT Line detection is the basis of the image segmentation, and also is one of the most important parts of image processing. It has many important applications on road identification, building recognition, medical image analysis. In the image processing field,line features are often used as a high-level processing,such as stereo matching, target tracking, object recognition. So the line detection and analysis for digital image processing have an important significance.  Straight line, especially segment, has simple geometric properties. Using them to describe the edge characteristics of target can be easy to classify regular objects,and improve the efficiency of subsequent image processing. While a number of researchers were devoted to work in the study of line detection,and also developed a number of algorithms, such as chain code detection,DOG function, least squares method, radon transform and so on, but the most mature and most widely used method is still the Hough transform at present. Hough transform was proposed by Paul Hough in 1962, after that, classical Hough transform had been improved by many people.It maks the Hough transform get a better use. Hough transform has many advantages that the robustness is good, the extraction is stable, the precision is high, the anti-chirp nature is strong. Therefore,it not only can recognize simple regular geometric graph in the background, but also can det