为解决高速公路和城市道路上复杂条件下的弱线漏检问题，提出了一种基于梯度增强和逆透视验证的车道线检测方法。该方法使用车道线的结构和对比度特征提取车道线区域，利用提取的车道线区域进行车道线和道路样本的选择，并采用基于模糊线性鉴别分析获得从彩色RGB图像到灰度图像变换的最佳投影系数，以确保车道线和道路像素间的灰度差异最大，从而有效突出道路上的弱线；利用逆透视变换对候选车道线间的空间位置关系进一步验证，以此找回漏检的虚线。不同场景、不同天气状况下的实际道路图像的实验表明，方法具有很好的鲁棒性和准确性。%For solving the problem of the missing detection of faint lanes in highway and urban road, a lane marker detection method based on gradient enhancing and inverse perspective mapping ( IPM) is proposed in this paper. This method first extracts potential lane markers based on the structure and contrast features of the lane marker and then selects lane marker and road samples from the extracted potential lane markers. The fuzzy linear discriminat a-nalysis is applied to obtain the most discriminative transformation coefficient from RGB color image to gray image, so that in the IPM image, the intensity difference between lane and road pixels is enlarged, which effectively en-hances the faint lane. In order to resolve missing detection of dotted lane, the IPM is further applied, and geometry relationship of potential lanes in IPM is tested and verified for removing false lanes and avoiding missing dotted lanes. Experiments on road images in different scenarios and different weather conditions demonstrate the robustness and accuracy of the proposed method.
展开▼
