POST-PROCESSING METHOD AND DEVICE FOR DETECTING LANES TO PLAN THE DRIVE PATH OF AUTONOMOUS VEHICLE BY USING SEGMENTATION SCORE MAP AND CLUSTERING MAP

摘要

A post-processing method for detecting a lane for planning a drive path of an autonomous vehicle by using a segmentation score map and a clustering map, wherein a computing device includes: the segmentation score map and the segmentation score map generated by a convolutional neural network (CNN); obtaining the clustering map; cause the post-processing module to detect a lane element including pixels forming the lane with reference to the segmentation score map, and generate seed information with reference to the lane element, the segmentation score map, and the clustering map; to do; causing the post-processing module to generate a base model with reference to the seed information, and to generate a lane anchor with reference to this; causing the post processing module to generate a lane blob with reference to the lane anchor; and causing the post-processing module to detect a lane candidate group by referring to the lane blobs and to generate a lane model by applying a line-fitting operation to the lane candidate group, thereby detecting the lane; A method is provided comprising.