FAST Pre-Filtering-Based Real Time Road Sign Detection for Low-Cost Vehicle Localization

Abstract

In order to overcome the limitations of GNSS/INS and to keep the cost affordable for mass-produced vehicles, a precise localization system fusing the estimated vehicle positions from low-cost GNSS/INS and low-cost perception sensors is being developed. For vehicle position estimation, a perception sensor detects a road facility and uses it as a landmark. For this localization system, this paper proposes a method to detect a road sign as a landmark using a monocular camera whose cost is relatively low compared to other perception sensors. Since the inside pattern and aspect ratio of a road sign are various, the proposed method is based on the part-based approach that detects corners and combines them to detect a road sign. While the recall, precision, and processing time of the state of the art detector based on a convolutional neural network are 99.63%, 98.16%, and 4802 ms respectively, the recall, precision, and processing time of the proposed method are 97.48%, 98.78%, and 66.7 ms, respectively. The detection performance of the proposed method is as good as that of the state of the art detector and its processing time is drastically reduced to be applicable for an embedded system.

Detection-Verification-Cascade-Based Low-Cost Precise Vehicle Positioning Exploiting Extended Digital Map

“…Therefore, a part-based detection method which extracts the corner points of a rectangle and combines them to detect a whole body is more applicable than the method utilizing the holistic feature [54][55][56][57]. Especially, road signs are generally installed high to make them easily observable and to prevent them from being occluded by other vehicles [54]. Moreover, a milestone among road signs is usually installed a few of kilometers before a junction (JC) or an interchange (IC) where the precise vehicle position is necessary for controlling a vehicle following the planed path.…”

A Precise and Robust Segmentation-Based Lidar Localization System for Automated Urban Driving

“…The line and plane equations can be expressed with a direction vector d d x , d y , d z or normal vector n n x , n y , n z and a point in it, where the point can be obtained by nearest neighbor search, and the key and difficulty is the solution of the vector. In the frame-frame matching of [17] and [28], they first find the two or three nearest points in different laser lines, then use Formula (5) and (6) to calculate the direction and normal vector, respectively.…”

“…The real-time carrier-phase based differential GNSS technology can eliminate satellite orbit and block errors, tropospheric and ionospheric delays to achieve centimeter-level localization accuracy, i.e. RTK (Real Time Kinematic) technology [5,6].…”

Low-Cost Precise Vehicle Localization Using Lane Endpoints and Road Signs for Highway Situations

“…According to information obtained by the perception sensor and contained in the digital map, localization methods utilizing a perception sensor with a digital map can be categorized into a feature point-based approach [5], [7]- [9] and a road facility-based approach [4], [6], [10], [11], [17]- [24], [41]. The feature point-based approach finds feature points using cameras and matches them with the corresponding feature points contained in the digital map.…”

“…Here, lane endpoints are used to enhance the longitudinal position accuracy and road signs are used to improve the ego-lane identification accuracy. Also, the landmarks used in the proposed system are detected by our previously developed methods: the lane and lane endpoint detection method [12], and the road sign detection method [17]. This system is ready to mount on massproduced vehicles since it only uses low-cost sensors such as a monocular front-view camera, in-vehicle wheel speed and yaw rate sensors, and a low-end global positioning system (GPS).…”