CP-loss: Connectivity-preserving Loss for Road Curb Detection in Autonomous Driving with Aerial Images

Abstract

Road curb detection is important for autonomous driving. It can be used to determine road boundaries to constrain vehicles on roads, so that potential accidents could be avoided. Most of the current methods detect road curbs online using vehicle-mounted sensors, such as cameras or 3-D Lidars. However, these methods usually suffer from severe occlusion issues. Especially in highly-dynamic traffic environments, most of the field of view is occupied by dynamic objects. To alleviate this issue, we detect road curbs offline using high-resolution aerial images in this paper. Moreover, the detected road curbs can be used to create high-definition (HD) maps for autonomous vehicles. Specifically, we first predict the pixel-wise segmentation map of road curbs, and then conduct a series of post-processing steps to extract the graph structure of road curbs. To tackle the disconnectivity issue in the segmentation maps, we propose an innovative connectivity-preserving loss (CP-loss) to improve the segmentation performance. The experimental results on a public dataset demonstrate the effectiveness of our proposed loss function. This paper is accompanied with a demonstration video and a supplementary document, which are available at https://sites.google.com/view/cp-loss.

csBoundary: City-scale Road-boundary Detection in Aerial Images for High-definition Maps

“…Few past works have exactly the same scope as this paper (i.e., automatically create the city-scale HD map of road boundaries from BEV aerial images), while they focus on related tasks, such as road laneline detection [5], [6], road network detection [7]- [10], road curb detection [11], [12] and road boundary detection [4], [13]. These works could be classified into three primary categories: segmentation-based methods, iterative-graphgrowing methods, and graph-generation methods.…”

“…They first predict the segmentation map of the target object and conduct post-processing algorithms to extract the final graph, such as skeletonization. Due to the poor topology correctness of segmentation-based methods, some recent works [7], [8], [11], [12] iteratively grow the graph vertex by vertex in a sequential manner. Even though this category of methods presents much better topology correctness, they suffer from the error accumulation issue and awful parallelization capability.…”