基于阈值三方向改进Hough变换的红外图像边缘检测

摘要

Aimed at the problem of the Hough transformation which is inability to deal. A new method propose to improvement the examination process, first infrared image Hough transform, compute all element segment (θ,p), the gentle vertical direction qnantifieation,then sampling point (i, j) goal element computation ladder to make optimization threshold, finally carries on three detections filter, use the 3 x 3 filter template realization, given the filter formula,the directions cover (0,2π). The experimental simulation through to the base chart, the simple chart, the complex chart examination withdraws its edge line segment characteristic to carry on the contrast; this article examination quality is clear,links, and eliminated artifacts.%针对Hough变换存在无法处理伪像的问题.提出了一种新的改进检测过程,首先对红外图像Hough变换,计算像素连线所有的(θ,ρ),对水平和垂直方向量化,然后对采样点(i,j)目标像素计算梯确定最佳阈值,最后进行三方向检测滤波,用3×3的滤波模板实现,给出了滤波公式,检测方向覆盖了(O,2π).实验仿真通过对基本图、简单图、复杂图检测提取其边缘线段特征进行对比,本文的检测质量较清晰、连贯,并消除了伪像.