METHOD AND DEVICE FOR DETECTING LANE ELEMENTS TO PLAN THE DRIVE PATH OF AUTONOMOUS VEHICLE BY USING A HORIZONTAL FILTER MASK, WHEREIN THE LANE ELEMENTS ARE UNIT REGIONS INCLUDING PIXELS OF LANES IN AN INPUT IMAGE

摘要

In a method of detecting a lane element, which is a unit region including a pixel of a lane in an input image, for planning a driving path of an autonomous vehicle using a lateral filter mask, a convolutional neural network (CNN) uses the input image. Generating a segmentation score map, obtaining, by a computing device, the segmentation score map; Causing, by the computing device, a post processing module that performs data processing at an output terminal of the CNN to generate a Magnitude map using the segmentation score map and the lateral filter mask; Causing, by the computing device, the post-processing module to determine each lane element candidate group for each row of the segmentation score map by referring to the value of the magnitude map; And causing the computing device to detect each lane element by causing the post processing module to apply a prediction operation to each of the lane element candidate groups for each row. Is initiated.