一种新的基于Hough变换的直线识别方法

A New Hough Transform Based Line Recognition Method

针对Hough变换识别直线时计算量大、准确率不高等问题,对斜率分式查表法Hough变换进行了改进,构建梯度阈值,对像素点进行分类,通过设定梯度阈值过滤掉不属于直线的像素点,解决Hough变换计算量大的缺点。并将最小二乘法与改进Hough变换相结合,先对数据点进行Hough变换,找出拟合直线,再采用最小二乘法对数据点进行直线拟合,来提高识别精度,这样既克服了直接用Hough变换识别时精度不高的缺点,又克服了最小二乘法在拟合直线时容易受噪声点干扰的缺点。

When the current calculation of straight line Hough Transform is larger,the accuracy is not high,the table slope of the fraction is improved for Hough transform to construct the gradient threshold,the classification of pixels,setting the gradient threshold filtering straight out of the pixel is to solve the shortcomings of large computing Hough transform;the least square method and the combination of this improved Hough transform,the first data point on the Hough transform is to find fitting line,and then the data using the least square method point is to fit a straight line to improve recognition accuracy,so that both the direct use of Hough transform is to overcome the identified shortcomings when accuracy is not high,and to overcome the least shortcomings of squares fitting for a straight line at the point.Experiments show that the slope of image recognition algorithm meets the requirements of practical application.