基于中介理论的车道线识别方法研究

以中介真值程度的数值化度量为基础,利用单个像素点的真值度,设计了新的图像去噪滤波算法。与经典算法相比,新算法具有较好的视觉效果。通过研究道路图像车道线的特点,采用中介距离比率函数来衡量像素点间相似程度,并设计了相应的图像边缘检测算法。与传统的道路图像边缘检测算法相比,新方法边缘增强效果显著。最后,基于道路直线模型,采用Hough变换实现了车道线的识别。

A Study on Lane Mark Detection Based on Medium Theory

A new image denoising algorithm using true degree of single pixel value is presented based on the numerical measure of medium truth scale.Compared with classical algorithm,the new algorithm has better visual effect.After analyzing the features of lane marks in road image,the similarity degree of pixels is weighed by medium distance ratio function,and a corresponding image edge detection algorithm is designed.The new method has better edge enhancement effect than traditional road image edge detection algorithm.Finally,by using Hough transform,lane mark detection is achieved based on straight-line road model.