A coarse-to-precise matching strategy in navigation

In order to ensure that SAR scene matching aided navigation system can acquire the position errors and yawing errors simultaneously, it is important to select the matching method carefully. In this paper, a coarse-to-precise matching strategy is proposed to be applied in navigation. The proposed method combines area-based and feature-based matching method. Specifically, in the coarse phase, a wavelet-based pyramid multi-resolution technique is adopted to reduce the number of calculated pixels, and then the robustness of the strategy is achieved by a multi-scale circular template fusion. In the precise stage, a nonlinear scale space method is recommended to accomplish the high matching performance. Two experimental results, matching real-time images with reference image and matching multiple real-time subimages with reference image, are listed to validate the proposed strategy.