Abstract
One of the main advantages of remote sensing technique is providing low coast and easy access to spatial data. Nevertheless, photogrammetric base map production process became easier and rapid due to developments of high resolution digital aerial cameras and computer systems. Therefore object extraction and integration to the GIS started to play very important role. Updated road data is required for transportation, navigation, environmental and urban planning issues. Automatic extraction of roads from satellite images is still popular subject for many researchers. However, it has many challenging problems as shadows of trees and buildings, cars, illumination effects. One of the major component of the road extraction is sensor resolution. Although only main axis of road can be extracted from low-resolution images, high resolution images can make possible to road borders too. Several methods have been proposed for road extraction by researchers. Morphological Analysis, Dynamic Programming and Snake, Segmentation, Vectorization, Automatic Seeding, Classification, Edge Detection, Hough Transform, Mathematical Morphology and Filtrate, Semantic Nets and Neural Networks, Filter based multi-scale image representation, Fuzzy logic methods, Fusion techniques can be given as example. In this study, Worldview-2 satellite images of Catalca and Silivri regions of Istanbul-Turkey have been used. Java based ImageJ open source software has been used as basic platform. Our developed method and related plugins have been integrated to the open source ImageJ software. The main steps of the study consist of noise elimination by median filter, applying morphological operators and Mahalonobis distance method for segmentation and road line extraction phases. Obtained results have been compared with base maps in the scale of 1: 5000. In the future, the developed algorithm will be integrated to the unmanned aerial vehicle images due to their increasing reliability and using possibility.
| Original language | English |
|---|---|
| Publication status | Published - 2015 |
| Event | 36th Asian Conference on Remote Sensing: Fostering Resilient Growth in Asia, ACRS 2015 - Quezon City, Metro Manila, Philippines Duration: 24 Oct 2015 → 28 Oct 2015 |
Conference
| Conference | 36th Asian Conference on Remote Sensing: Fostering Resilient Growth in Asia, ACRS 2015 |
|---|---|
| Country/Territory | Philippines |
| City | Quezon City, Metro Manila |
| Period | 24/10/15 → 28/10/15 |
Keywords
- Automatic road extraction
- Image processing
- Mahalonobis distance
- Morphology
