科研之友
微信
新浪微博
Facebook
分享链接
摘要
Although road geometry data can be automatically collected using instruments mounted on survey vehicle, measurement of curved ramp geometry is still of low effectiveness and accuracy due to manual or semi-automatic detection of PC (Point of Curvature)/PT (Point of Tangent) as well as influences of vehicle vibration and wandering. In this study a new method is presented for automatic measurement of ramp geometry in network level using IMU (Inertial Measurement Unit) and 3D LiDAR (Light Detection And Ranging) system. Firstly, horizontal alignment measurements are implemented: 1) an improved K-Mean clustering method and linear fitting method are integrated for automatic PC/PT station detection; 2) an algorithm is developed for automatic lane marking identification and localization for vehicles trajectory calibration; 3) curve radius and length are measured based on roadway centerline. Subsequently, pavement slope is calibrated using IMU and transverse profiling data. Finally, nine segments are chosen from highway ramps as test bed, and validation tests are conducted using the field measurement. The test results show the average errors for curve detection and curve radius measurement are 5.89% and 1.99% respectively, and the P-value for longitudinal and cross slope measurement are 0.621 and 0.989 respectively, which indicate the proposed method is robust in ramp geometry measurement. The significant of the proposed method is three folds. First, it integrates and synchronizes the IMU and 3D LiDAR system in geometry measurement. Second, it solves the common problems of mobile survey on vehicle wandering and vibration. Third, it is of high accuracy and effectiveness, and can be used for roadway survey in network level.
- 出版日期2018-10
- 单位福建农林大学
