To address the problems of complex algorithm structure and large number of parameters existing in road lane detection, a lane-detection method based on dual-branch feature extraction (DP-RESA) is proposed. This model treats road-lane detection as a semantic segmentation task. First, lane features were extracted using two parallel branches: a semantic branch and a detail branch. The semantic branch uses a lightweight MobileNetV2 model to extract high-level features, while the detail branch uses wide channels and shallow layers to capture low-level details with more spatial details. Furthermore, the model quickly and efficiently fuses features by using high-level features as weights to filter essential semantic information embedded within the low-level features. Finally, the experimental results for the Tusimple dataset show that compared with the baseline, DP-RESA network lane detection accuracy can reach 96.58%. In addition, the number of model parameters is reduced to 5.12 MB, and the single-image inference time is reduced by 11.76 ms, making the model well suited for lane-detection tasks deployed on resource-constrained embedded platforms.