【毕业论文设计】基于Hough变换的圆检测算法研究及其实现.docVIP

基于Hough变换的圆检测算法 研究及其实现 The Study and Realization of Circle Detection Based on Hough Transform 摘 要 图像是人类获取和交换信息的主要来源。圆是构成图像的一个基本要素，因此，能够快速、准确地对圆进行检测，在图像处理中显得尤为重要。 Hough变换是一种提取直线、圆、椭圆、二次曲线甚至是任意形状边缘的有效方法，目前已经在军事和民用领域将会得到广泛的应用，如：图像处理、信号检测、雷达目标跟踪、被动跟踪、多传感器多目标跟踪等。但是，Hough变换大多数算法的计算量大，需要很大的存储空间，而且都是假设图像在计算机中能用完美的模型来描绘。然而，由于噪声、数字化误差等因素影响，真实的图形在计算机中经常会失真。 本文针对Hough变换计算量大的情况研究了一种新的基于Hough变换的圆检测的快速算法，即先检测出一个内、外半径已知的圆环，然后在整个图像范围内移动这个圆环，对移动后在圆环范围内的点的个数进行累加，累加得到的局部最大值提供了检测出的圆心坐标。这种方法实现了各类圆形、椭圆等非理想图形的快速检测，克服了Hough变换耗时巨大的缺陷。最后，本文采用VC++编程实现了这种Hough变换的圆检测并进行了实验研究，实验结果表明，该快速算法能够对各类圆形、椭圆等非理想图形进行快速检测，有一定的抗噪声性能。该算法没有很好的改善Hough变换需要很大的存储空间的问题，这将是今后研究的一个方向。 关键字：图像处理，圆检测，Hough变换,快速算法 Abstract Image is the main source that people gain and exchange information, the application domain of image processing must relate to human life and work in all its aspects. The circle is a basic essential factor constituting the image. So it is important that detecting the circle quickly and accurately in image processing especial the industrial image processing. Hough transform is an effective way to detect line、circle、ellipse even any shape . For many years, researchers all devote to study the algorithm of Hough transform. Now Hough transform has been widely used in military and civil domain, such as image processing, signal detection, radar target track, passive track, multitarget-multisensor track and so on. But most of the algorithms of Hough transform cost much time and memory space. And they all suppose that the image can be well described by perfect models in computers. Otherwise, in our real life, real images are distorted because of the influence of noise, digital error and so on. Although the image input is complete, the image in computers is difficult to be well described by perfect models. Aimed at the problem of huge time consuming, this paper studied a fast algorithm for the detection of circles. The principle is that we first detect a cirque which has a definite inside and outside radius,