Extracting accurate ground control points in desert areas by matching between aerial and satellite images

Abstract EN

This paper came out to solve a critical problem which appeared during using SIFT method and other similar existing methods, the problem involved the difficulty of extracting ground control point in specified areas, mainly, deserts areas.

The Ministry of Municipal and Rural Affairs in Saudi Arabia requested a solution for this problem, particularly because most of the Saudi Arabia areas are deserts.

All of the existing proposed methods failed to handle the problem to extract ground control points in deserts areas.

SIFT method is the standard method for the extraction of key points.

SIFT, as it is found now, is not used completely and does not fit with the deserts areas, the description of the key points was the main modification made to solve the problem.

This modified method presents high quality orthorectified products and geometric accuracies of approximately one pixel size.

To guarantee that the automatically extracted ground control points from the aerial photos provided by this modified method are on a high level of accuracy, three samples from the data set satellite and aerial images were taken from three different locations in Saudi Arabia, the samples were distributed as: • Deserts without buildings.

• Deserts which include buildings within it.

• Urban areas.