Morphological Correlator Based on Laplace Filtering

As a modified approach to improve the properties of the classical Morphological Correlation(MC),a Laplace Filtering-based Morphological Correlator(LBMC) is proposed for pattern recognition.Different from classical MC,for the LBMC,the Joint Power Spectrum(JPS) summation of the MC is filtered by power spectrum of Laplace operator before final inverse Fourier transform.Computer simulation results show that,compared with the Linear Correlation(LC) and the conventional MC,LBMC provides better discrimination capability with sharp and strong unmistakable correlation signal,and its ACR(autocorrelation peak intensity to cross correlation peak intensity ratio) and Peak to Noise Ratio(PNR) are more stable when input scene is corrupted by outlier noise(salt-and-pepper noise).Although the LBMC is not illumination invariant when multiplication illumination factor is larger than unity,considerable discrimination capability is still obtained.