A new ground control point image-chip automatic matching method for remote sensing image location processing

摘要

Get access to ground control points(GCPs) by ground control point image-chip matching is the key to automated precise geometric correction processing which is a great improvement compared to artificial reference map or the GPS method. The key to obtain GCPs is the correct matching between GCP image chips and images to be corrected. The existing gray-scale based normalized correlation and edge-based normalized matching algorithms are in-depth studied, and a method that combined normalized correlation and Renyi entropy mutual information multi-criteria image matching is proposed. Experiments show that the new method did well on matching homologous and heterologous images and will be a good support for GCP quick acquisition and automated precise geometric correction processing.