Experiments using microarray technology generate large amounts of image data that are used in the analysis of genetic function. An important stage in the analysis is the determination of relative intensities of spots on the images generated. This paper presents GridWeaver, a program that reads in images from a microarray experiment, automatically locates subgrids and spots in the images, and then determines the spot intensities needed in the analysis of gene function. Automatic gridding is performed by running Fast Fourier Transforms on pixel intensity sums. Tests on several data sets show that the program responds well even on images that have significant noise, both random and systemic.