Adaptive Balancing of Edge Extraction in LADAR-Referenced Navigation over Plain Area

Although high-resolution remote sensing (RS) images can supplement the reference map´s deficiency in LADAR (Laser Radar) referenced navigation system over plain area, their different imaging modes still leave great errors´ introduction into the positioning process. As positioning is commonly accomplished by matching the edges extracted as the obvious terrain characteristics, exploring automatic procedure for balancing mutual information is an important step to improve the system´s performance. To solve this issue, this paper explores the criterion of edge-points ratio (EPR) based on wavelet multi-scale analysis extraction method, and then appropriate measures are applied for different conditions. EPR at various decomposition layer works as the key circle for adaptive balancing, and so related edge-distribution characteristics can be used individually, such as close-contour or line.