| Landmines contain all forms of explosive materials. Many landmine detection techniques that already exist have different complexities. The efficient solution for the problems created by landmines means that close to hundred percent of mines in any area must be identified at the fastest rate with few false alarms. Hence, advanced signal processing and signal fusion methods are required.;This thesis aimed to evaluate the gain in performance when nonlinear digital fusion approaches are used. Discrete wavelet transform (DWT), laplacian pyramid (LP) algorithms and logical operators were used to fuse images obtained by a polarization sensitive sensor. Both the raw and the fused images were then enhanced using high frequency emphasis (HFE).;Fusion of images has not shown improvement in the detection capability due to the failure of enhancement of edge features but the application of HFE on the images has provided improvement in the detection capability. The concept of wavelet filter joint transform correlation (WF-JTC) was used and the improvements were justified using several correlation performance metrics (CPM). The gain in the performance of each method was evaluated using the values of peak-to-sidelobe-ratio (PSR). |
