基于角点检测的图像处理方法大学本科毕业论文.docVIP

基于角点检测的图像处理方法研究 摘 要：本文主要研究了图像的角点检测方法，在计算机视觉中，机器视觉和图像处理后总，特征提取都是一个重要的方向。而角点又是图像的一个重要局部特征，它决定了图像中目标的形状，因此在图像匹配，目标描述与识别及运动估计，目标跟踪等领域，角点提取都具有重要的意义。角点的信息含量很高，可以对图像处理提供足够的约束，减少运算量，极大地提高运算速度。角点检测问题是图像处理领域的一个基础问题,是低层次图像处理的一个重要方法。角点检测的目的是为了匹配，而匹配的效率取决于角点的数量。Harris角点检测原理是对于一副图像，角点于自相关函数的曲率特性有关，自相关函数描述了局部局部图像灰度的变化程度。在角点处，图像窗口的偏移将造成自相关函数（图像灰度的平均变化）的显著变化。harris算子是一种简单的点特征提取算子，这种算子受信号处理中自相关函数的启发，给出与自相关函数相联系的矩阵M。M阵的特征值是自相关函数的一个阶曲率，如果两个曲率值都高，那么久认为该点是特征点。 关键词：角点，角点检测，Harris角点 ABSTRACT This paper studies the image of the corner detection methods in computer vision, machine vision and image processing general, feature extraction is an important direction. The corner is an important local feature image, which determines the shape of the target image, so the image matching, object description and recognition and motion estimation, target tracking and other fields, corner detection are of great significance. Corner of the information content is high, image processing can provide sufficient constraints to reduce the amount of computation greatly improve the processing speed. Corner detection is a basic image processing problems, low-level image processing is an important way. Corner detection is designed to match the efficiency of the matching depends on the number of corners。Harris corner detection principle is that for an image, corner point on the curvature properties of the autocorrelation function is related to the local auto-correlation function describes the degree of local image intensity changes. In the corner point, the offset will result in the image window autocorrelation function (the average image intensity changes) change significantly. arris operator is a simple point feature extraction operator, this operator by the signal processing in the autocorrelation function of inspiration, given the autocorrelation function associated with the matrix M. Eigenvalues ??of matrix M is an order autocorrelation function of the curvature, if the two curvature values ??are high, for so l