Arata Takahashi

Model-based lane recognition

A lane recognition is used in various driver assist systems. The lane recognition detects lane boundaries and gets vehicle position relative to the lane and lane structure. A vision system is good for lane recognition because the vision can detect lane marks. One of the subjects for the vision system is improvement of robustness. Various methods have been tried to achieve it. We tried to improve the model-based approach for the robustness. Our main idea is the noise reduction based on narrowing a width of search area. The proposed method uses the road model based on the space continuity of lane structure. An update of the model is calculated through extended Kalman filter based on lane structure restriction. A search area width is estimated from covariances of the model parameters. After model parameters and their covariances are updated from observed lane boundaries near the vehicle, search area width for distant lane boundaries becomes narrower than the previous update. After the several updates, distant lane boundaries can be detected robustly with narrow search area excluding noisy image features.

Rear view lane detection by wide angle camera

In this paper a new lane detection method is proposed for a highway lane departure warning system which hardware cost is reduced by using an existing wide-angle back monitor camera and an embedded CPU on the market. The extended Hough transform and a birds-eye view image are effectively applied to perform realtime lane detection. The method was built on an embedded CPU (Hitachi SH2E) as software. Then the extended Hough transform was executed at less than 66 milliseconds.

A robust lane detection using real-time voting processor

We describe a lane detection method using RVP-I (Real-time Voting Processor-I) for a driver assistance system and an auto driving system. For robust lane detection in various environments, we propose a method based on complete search in a parameter space. The performance of the proposed method has been investigated on 1000 road images of bad conditions. A 95% detection rate is obtained in comparison with 81% of our previous method. Thus the proposed method is effective for robustness of lane detection.

Accurate Vehicle Localization using DTW between Range Data Map and Laser Scanner Data Sequences

We propose a method for accurate vehicle localization. The proposed method detects a vehicles location and traveling lane by matching between a pre-constructed range data map and laser scanner data sequences measured while the vehicle runs. The range data map consists of an absolute position on the road and range data at the position. We use dynamic time warping (DTW) to align multiple range data sequences. Experiments using 40 data sequences collected while a vehicle ran on the same route with multiple traffic lanes were conducted. The results demonstrated the effectiveness of vehicle localization and traveling lane classification.

A method for robust lane marking candidate selection of lane recognition

Similar lane marking patterns are captured in the images of rainy roads because of surface reflections, resulting in false selection of lane marking candidates and giving rise to problems due to estimation errors. The authors applied lane structure heuristics, a method of improving robustness by reducing false selection of similar patterns, to lane marking candidate selection. The effectiveness of the proposed method was confirmed by applying the method to experiments with real images of a rainy highway and comparing the results with those obtained by LMedS estimation.