FPGA Based Parallel Architectures for Normalized Cross-Correlation

As a similarity measure, normalized cross-correlation has found application in a broad range of image processing. A dedicated hardware implementation of normalized cross-correlation is crucial for the requirements of real-time high-speed tasks such as automatic target matching, recognition and tracking. Two efficient parallel architectures for real-time implementation of normalized cross-correlation using field programmable gate array (FPGA) are proposed in this paper. In these architectures, several novel efficient approaches are proposed to reduce logic resource usage and computation time. These two architectures can be applied in different situations according to the practical available resource of the FPGA chip used. Function and timing simulation with Quartus II 8.0 and practical experiment in target recognition using image matching have shown that these architectures can effectively improve the speed performance of the practical target recognition system.