Investigation on Parallel Computing Techniques for Multiple-Image Matching AMMGC Model

Stimulated by sensor technology, image matching techniques are greatly innovated towards multiple-image matching. In this paper, an image-space-based AMMGC multiple-image matching model is introduced. However, AMMGC model is quite complex and involves massive computing amount, especially for dense image grid point. So, two parallel computing methods are analyzed comprehensively from the point of average image data partition. The methods can be greatly integrated with AMMGC model to provide an easy matching procedure. Experimental results prove that AMMGC model has quite reliable matching quality, and can be greatly combined with parallel computing techniques to reduce matching time, and at the same time, enormously improve the matching speed and scale.