Abstract:
Remote Sensing images are consists of photographs of Earth or other planets captured by means of satellites, helicopter,
rocket, drone etc.. The quality of remote sensing images depends on sensor, camera used to capture images
and number of bands. Due to repaid development of technologies made possible to access very high resolution
remote sensing images through Quick Bird, Ikonos and many more sources. The applications of high resolution
remote sensing images mainly in agriculture, geology, forestry, regional planning, geographic map updating and in
the military. Extensive investigation has been proposed to detect road features from remote sensing images. Roads
are the backbone and essential modes of transportation, providing many different supports for human civilization.
The research of road extraction is of great significance for traffic management, city planning, road monitoring,
GPS navigation and map updating. To identify and distinguish roads elements from remote sensing images which
have similar spectral characteristics type background objects like buildings, rivers, and trees is a challenging task.
This paper presents a summary of various road network detection methods from Remote Sensing (RS) images with
respect to resolution of test and training images, accuracy, road features, advantages and limitation of method. It
also gives information about recent approaches to extract road network from remote sensing images.
