中文摘要：

为了能有效解决Hough变换的计算量大的问题，文中提出了一种基于直线局部结构特征的Hough变换改进的直线检测算法。该算法根据Freeman准则分析了直线上基元的特征信息，通过图像上邻近的同类基元的倾斜角约束基元上像素点的极角范围，减少每个点的计算次数，在保持精度的同时，提高直线检测的速度，在有噪声的情况下，该算法相对标准Hough变换算法可以提高到6到7倍。

英文摘要：

In order to effectively solve the problem of decreasing computation in the process of using Hough Transform（HT） to extract straight line, an improved HT algorithm methocl is presented based on the local characteristics of the straight line. According to Freeman Criteria, the improved method analyzes elementary line segments structure of the straight line, and defines the scope of polar angle of the points on elementary line segments by computing the tilt angle of the straight line determined by the two adjacent and congeneric elementary sine segments,and then decreases the computing times of every point. Pointing to the image containing noise, the testing data shows that velocity of the improved algorithm is 6-7 times as fast as the standard one,while the precision is kept.