Lane line detection method, apparatus, and computer-readable storage medium

A lane line detection method, apparatus, and compute readable storage medium are disclosed. The lane line detection method comprises the following steps of: training a neural network for detecting a small block of a lane line; Inputting an acquired image to a trained neural network, and detecting a plurality of lane line patches in the acquired image through the neural network, wherein a portion of the lane line patches is included in each lane line patch; Extracting a corresponding lane line feature point based on the detected lane line small block; And based on the piecewise fitting model, the feature points of the extracted lane line are fitted to obtain the lane line. 公开了车道线检测方法、设备和计算机可读存储介质。所述车道线检测方法，包括：训练用于检测车道线小块的神经网络；将采集图像输入至完成训练的神经网络，并通过所述神经网络检测所述采集图像中的多个车道线小块，其中在每个车道线小块中包括部分车道线；基于检测到的所述车道线小块，提取对应的车道线特征点；以及基于分段拟合模型，对提取出的车道线特征点进行线拟合，以获得车道线。