Identification of Roads from CARTOSAT-2 Satellite Images Using Line Segment Detector

Abstract

Roads are the basic component for transportation. Roads play an important role in human life these days. If the navigation is accurate, the transportation will be even more easy for the humans. Every year new roads are planned and they are infornt of humans. New roads data will be taken by the satellite and updates to map. Identification of roads information from a satellite image or from a map is easy for humans but it is difficult for a computer. The images given by the CARTOSAT-2 satellite is panchromatic images. These images are usually gray-scale images. Extracting information form panchromatic images has been difficult task for the computer due to the darkness of the image. In this paper, we work for identificating roads information from the panchromatic images to reduce the difficulty to complete the task by the computer. The CARTOSAT-2 satellite data is available as per the request by the user in BHUVAN website. This site helps to provide satellite data to users those who wants to do research on images and those who wants to add missing data manually and to update the data and etc. In this attempt, we have applied digital image processing techniques on the panchromatic images which leads us to identify roads imformation approximately from the panchromatic satellite images. The proposed algorithms from image processing techniques are helpful to identify data. The identified roads information is useful for the creation of maps and even in different implementations such as urban and rural planning, disaster and utility management, cartographic application and various other Geographical Information Systems (GIS).