基于多尺度复合卷积和图像分割融合的车道线检测算法

Lane Line Detection Algorithm Based on Multi-Scale Composite Convolution and Image Segmentation Fusion

在车道线检测任务中,由于车道线的特点和获取更大范围感受野的需求,空洞卷积被广泛使用.然而,为了获取大范围信息,空洞卷积会造成卷积点附近信息的丢失.针对以上问题,提出了一种基于多尺度复合卷积和图像分割融合的车道线检测算法.首先将不同尺寸的空洞卷积、全卷积和标准卷积结合以弥补空洞卷积造成的信息丢失;然后通过语义分割和实例分割融合的图像分割融合模块来增强实例分割网络对全局特征的关注;最后,设计一个加权交叉熵损失函数对网络进行训练和优化.实验结果表明,算法在CULane数据集中的整体F1measure取得74.9%,整体性能优于比较算法,在多种挑战性环境中均有所提升.

In the task of lane line detection,atrous convolution is widely used due to the characteristics of lane lines and the need to obtain a wider receptive field.However,in order to obtain large-scale information,atrous convolution will cause the loss of information near the convolution point.To solve the problems,a lane line de-tection algorithm was proposed based on multi-scale composite convolution and image segmentation fusion.First,atrous convolution,full convolution and standard convolution of different sizes were combined to com-pensate for the loss of information caused by atrous convolution.And then,fusing semantic segmentation and in-stance segmentation,an image segmentation fusion module was arranged to enhance the attention of instance segmentation network to global features.Finally,a weighted cross-entropy loss function was designed to train and optimize the network.The experimental results show that the overall F1measure of the algorithm in the CULane dataset can achieves 74.9%.Comparing with other algorithms,the overall performance of the proposed detection algorithm is better and improved in various challenging environments.