Lane marking identification is one of the key technologies in an intelligent vehicle vision system. To effectively segment the lane marking is significant for the recognition of the lane marking. Based on the target of the lane segmentation, combined with the characteristics of the road images, the interference effects of noise on the segmentation are studied, and an improved image segmentation algorithm based on the Otsu method is presented. The algorithm is used for each line of image to calculate the Otsu threshold value, and then the threshold is amended. Finally, four images are used to test the improved algorithm, and the results are compared with the results of the Otsu method. The experimental results show that the improved algorithm can remove more noise in lane marking segmentation, has robust, strong adaptive capacity and segmentation results are improved.