The topic of change detection of different topographic features on the field using aerial images or satellite high-resolution imagery in digital format is very important for the military domain as well as for the civilian one (agriculture, forestry, natural disaster management, etc.). In this paper the authors treat some methods used for automated detection of changes on the terrain based on panchromatic aerial photographs (gray levels, 8 bits for color representation) acquired at different dates. One method is implemented in ERDAS Imagine software and the others were developed empirically by a third party. A new method for change detection is proposed by the authors, using Sobel edge detection operator, obtaining thus clearer contours of modified elements on the ground. The quality of the resulting modified elements depends very much on technical characteristics of compared images such as collection dates, atmospheric conditions during the capturing process, similarity of radiometry, spatial resolution and precise overlay, as well as the presence of clouds or shadows of topographic details on the images.
Nowadays are widely used geospatial information-based software applications, among which navigation systems on vehicles, systems monitoring and management of traffic etc., requiring the use of accurate, complete and up-to-date roads, stored in different types of databases. There are many automated or semi-automated algorithms used for road extraction, but in this article is presented a new semi-automated algorithm for extracting roads from high resolution aerial and satellite images based on the weighted correlation of transverse profiles. The algorithm uses, as initial data, two starting points from which one obtains the path's orientation and template profile. Also, the operator must set threshold value of correlation coefficient between cross profiles, search distance, search angle, length of transverse profile and maximum number of rejections. Compared to other semi-automated road extraction algorithms, this algorithm is less sensitive to radiometric changes at the ends of the profile due to the assignment of higher weights to central pixels.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.