Image processing enhancement of high-resolution TEM micrographs of nanometer-size metal particles
The high-resolution TEM detectability of lattice fringes from metal particles supported on substrates is impeded by the substrate itself. Single value decomposition (SVD) and Fourier filtering (FFT) methods were applied to standard high resolution micrographs to enhance lattice resolution from particles as well as from crystalline substrates. SVD produced good results for one direction of fringes, and it can be implemented as a real-time process. Fourier methods are independent of azimuthal directions and allow separation of particle lattice planes from those pertaining to the substrate, which makes it feasible to detect possible substrate distortions produced by the supported particle. This method, on the other hand, is more elaborate, requires more computer time than SVD and is, therefore, less likely to be used in real-time image processing applications.