Skip to content

Satellite imagery semantic segmentation requires high-resolution satellite images to train the Nerual Network. Due to the expensive cost of the pixel-level annotation, the cheaper instance-level annotation of the satellite image is not suitable for the high-resolution image, which leads to the high error rate. Driven by the need to avoid a large…

Notifications You must be signed in to change notification settings

ShenweiXie/Stagewise-Weakly-Supervised-Satellite-Imagery-Road-Extraction-Based-on-Road-Centerline-Scribbles

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 

Repository files navigation

Stagewise Weakly Supervised Satellite Imagery Road Extraction Based on Road Centerline Scribbles

Extracting roads from satellite images through semantic segmentation algorithm has become the mainstream solution for Remote Sensing-based road monitoring tasks. However, due to complex features and changeable textures of roads in satellite imagery which derive from various geographical environments and the high cost of pixel-level road labeling, it is unaffordable to acquire a substantial dataset with pixel-level road annotation to train semantic segmentation models. To solve the above problems, a stagewise weakly supervised road extraction algorithm based on road centerline scribbles is proposed. The feature of road centerline scribbles is learned in a weakly supervised way, and the road segmentation model is trained by stages. In addition, the pseudo mask update strategy and the hybrid training strategy are proposed, and the loss functions for road foreground and road background are designed. The results show that compared with other weak supervision methods based on road centerline, the proposed algorithm achieves superior performance in road segmentation task. Ablation studies are also conducted to verify the effectiveness of the proposed training strategy.

@title = {Stagewise Weakly Supervised Satellite Imagery Road Extraction Based on Road Centerline Scribbles},  
@author = {Shenwei Xie (BUPT PRIS)}
@time = {from 2020/11/01}

0. Introduction and Related Papers


1. Code


2. Datasets and Benchmarks


About

Satellite imagery semantic segmentation requires high-resolution satellite images to train the Nerual Network. Due to the expensive cost of the pixel-level annotation, the cheaper instance-level annotation of the satellite image is not suitable for the high-resolution image, which leads to the high error rate. Driven by the need to avoid a large…

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published