# Enhancing Inverse Perspective Mapping for Automatic Vectorized Road Map Generation Source: [Enhancing Inverse Perspective Mapping for Automatic Vectorized Road Map Generation]() Collection index: [llms.txt]() Data updated: 2026-01-29T05:00:05.136Z Authors: Hongji Liu, Linwei Zheng, Yongjian Li, Mingkai Tang, Xiaoyang Yan, Ming Liu, Jun Ma Published: 2026-01-27T12:27:27.000Z ## Abstract In this study, we present a low-cost and unified framework for vectorized road mapping leveraging enhanced inverse perspective mapping (IPM). In this framework, Catmull-Rom splines are utilized to characterize lane lines, and all the other ground markings are depicted using polygons uniformly. The results from instance segmentation serve as references to refine the three-dimensional position of spline control points and polygon corner points. In conjunction with this process, the homography matrix of IPM and vehicle poses are optimized simultaneously. Our proposed framework significantly reduces the mapping errors associated with IPM. It also improves the accuracy of the initial IPM homography matrix and the predicted vehicle poses. Furthermore, it addresses the limitations imposed by the coplanarity assumption in IPM. These enhancements enable IPM to be effectively applied to vectorized road mapping, which serves a cost-effective solution with enhanced accuracy. In addition, our framework generalizes road map elements to include all common ground markings and lane lines. The proposed framework is evaluated in two different practical scenarios, and the test results show that our method can automatically generate high-precision maps with near-centimeter-level accuracy. Importantly, the optimized IPM matrix achieves an accuracy comparable to that of manual calibration, while the accuracy of vehicle poses is also significantly improved. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Lane line mapping on Sequence 1 Short \<500 m](): Lane line mapping; Sequence 1 Short \<500 m; APE (Lanes): 0.07. - [Lane line mapping on Sequence 2 Short \<500 m](): Lane line mapping; Sequence 2 Short \<500 m; APE (Lanes): 0.08. - [Lane line mapping on Sequence 3 Long \>1000 m](): Lane line mapping; Sequence 3 Long \>1000 m; APE (Lanes): 0.08. - [Lane line mapping on Sequence 4 Long \>1000 m](): Lane line mapping; Sequence 4 Long \>1000 m; APE (Lanes): 0.09. - [Mapping Accuracy Evaluation on Sequence 1](): Mapping Accuracy Evaluation; Sequence 1; Marking APE: 0.16. - [Mapping Accuracy Evaluation on Sequence 2](): Mapping Accuracy Evaluation; Sequence 2; Marking APE: 0.2. - [Mapping Accuracy Evaluation on Sequence 4](): Mapping Accuracy Evaluation; Sequence 4; Marking APE: 0.21. - [Mapping Accuracy Evaluation on Sequence 3](): Mapping Accuracy Evaluation; Sequence 3; Marking Accuracy (APE): 26. - [Lane Line Map Construction on Lane mapping 300m road segment](): Lane Line Map Construction; Lane mapping 300m road segment; Point Count: 96. - [Lane Line Map Construction on Lane mapping (1360m road segment)](): Lane Line Map Construction; Lane mapping (1360m road segment); Point Number: 453. ## Original sources and implementation [Original paper]() [PDF]()