期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII - Part 5
页码：381-386
出版社：Copernicus Publications
摘要：Image-based change detection concerns translation of two-dimensional intensity data into three-dimensional changes. The challenges that such analysis offers dictate in many cases imposition of constraints on the camera placement and illumination conditions. Evaluation of changes is then carried out either via point-to-point differencing or via segmentation and classification of the data. Overcoming these limitations, this paper proposes a three-dimensional image-based change detection model that integrates the stochastic nature of the observations into the analysis. In order to relax the need for fixed reference points that are marked in advance, a requirement that cannot always be fulfilled, the paper addresses also the registration of images from different epochs. For registration we propose a surface based alignment model and show that using this scheme, the registration and consequent change documentation can be viewed as two sides of the same problem. Change-detection is then carried out via outlier-detection analysis associated with geodetic network design and measurement. The paper discusses the application of proposed model for the documentation of an archeological excavation site. Results show the applicability of the model as well as its suitability for feature extraction
关键词：Close range photogrammetry; surfaces; modeling
