More Efficient Methods among Commonly Used Robust Estimation Methods for Similarity Transformation

Abstract

mage calibration is a necessary step in image processing. Similarity transformation is a widely used image calibration method. Robust estimation methods are often used to eliminate or weaken the influences of gross errors on image calibration. However, different robust estimation methods have different capabilities in eliminating or weakening gross errors. The current paper employed simulation experiments using different coincident points and the number of gross errors included in the observations to compare the robustness of 13 commonly used robust estimation methods. Results indicated that L1 and GermanMcClure methods are relatively more efficient than other robust estimation methods for image calibration based on similarity transformation.