In many industrial machine vision and environment sensing tasks, the acquisition and analysis of a single image is not sufficient because the information of interest cannot be captured with one image only. In such cases the desired information can often be obtained by combining a series of images to a result yielding a new or an enhanced representation of the image contents or characteristics. This contribution discusses essential aspects of image data fusion. First, image sensors are characterized with respect to their fusion-relevant features. Then, possible abstraction levels to accomplish the combination of images are described and a general method to perform the image fusion is presented. Finally, examples are given which demonstrate the performance of the fusion methodology presented.
Print ISSN: 0171-8096
Volume: 74, 03/2007
Pages: 130 - 138