Fig. 1. Structure of SCNNLane network

Fig. 2. Image preprocessing

Fig. 3. Overall architecture of VGG16-SCNN

Fig. 4. Structure of spatial convolution model

Fig. 5. Decode output images of lane line

Fig. 6. Structure of coupling network

Fig. 7. Decode output images

Fig. 8. Encoding partial visualization

Fig. 9. Output results of SCNNLane network after training

Fig. 10. Comparison of visualization results under Tusimple dataset