LineNet: a Zoomable CNN for Crowdsourced High Definition Maps Modeling in Urban Environments
This work addresses the problem of efficient HD map creation for urban traffic applications, though it appears incremental as it builds on existing lane detection methods.
The authors tackled the slow development of High Definition (HD) maps due to cost by proposing LineNet, a convolutional neural network with a novel prediction layer and zoom module for lane detection, and introduced the TTLane dataset, achieving precise HD map modeling from crowdsourced data for the first time.
High Definition (HD) maps play an important role in modern traffic scenes. However, the development of HD maps coverage grows slowly because of the cost limitation. To efficiently model HD maps, we proposed a convolutional neural network with a novel prediction layer and a zoom module, called LineNet. It is designed for state-of-the-art lane detection in an unordered crowdsourced image dataset. And we introduced TTLane, a dataset for efficient lane detection in urban road modeling applications. Combining LineNet and TTLane, we proposed a pipeline to model HD maps with crowdsourced data for the first time. And the maps can be constructed precisely even with inaccurate crowdsourced data.