SEM noise reduction for surface reconstruction. A high level of noise in an electron micrograph can adversely affect the final 3D geometry, resulting in a highly rough appearance (a). In order to minimize the effect of noise, a convolution algorithm can be applied. The surface presented in (b) was processed using a 3 × 3 convolution matrix to ‘smooth out’ the effect of noise. Care must be taken however, to ensure that the image is not excessively smoothed, as this will result in the loss of topographical detail. This is demonstrated in surface (c), where the same convolution matrix was applied 5 times in succession.