An Automatic Approach of Road Regions Extraction from Satellite Images based on Connected Component Algorithm

Satellite images provide valuable information for a wide range of applications. They use various spectral bands to recognize a great number of features. High-resolution satellite (HRS) images are able to represent man-made features such as roads and buildings. Roads are the most fundamental structures. We have presented a methodology using image processing and connected component method for automatic extraction of road network from any high resolution satellite image. It is sometimes challenging in the areas having complex scenes. There is an algorithm to extract the manmade features from the image and filtering operations are used to correctly extract out the respective feature i. e. the road network from the image. The proposed system follows various steps consisting of Enhancement, Filtering operations, Segmentation etc. The final extracted road network can be used for the map creation which shortens the human efforts in creating large scale maps and also it is very efficient in time and cost. The method is reliable to implement in real time also and can be applied for the real time development in the field of remote sensing and Geographic Information System including rural and urban development. The algorithm promises to provide good results and helpful in keeping Geographical Information System (GIS) updated with new road databases by reducing the time consumption in producing large scale maps.