The development of road information extraction from remote sensing images

Abstract

The automatic extraction of roads from high spatial resolution remote sensing images is the hotspot and difficulty in remote sensing research, and has attracted extensive attention. Road extraction from remote sensing images is a leading direction of remote sensing image processing and widely demanded in transportation, mapping, urban planning and other fields. In this paper, the actuality of road extraction is investigated. The road extraction process is divided into five phases, which are pre-processing, low-level processing, mid-level processing, high-level processing and the application of extraction results. The methods used in each phase are analyzed. Low-level processing is considered to be the key and foundation of road extraction. Based on the representative method in low-level processing, mathematical morphology, this paper presents an approach for automatic road extraction and tests the method on remote sensing image. The experimental result shows the efficiency of the presented approach.