Technologies For True Orthophoto Generation Based On Multi-angle And Multi-view Images

The orthophoto can be considered as an image on which effects of terrain height and camera attitudes are removed. It bears the features of high accuracy, rich information, and visual verisimilitude. Hence, DOM can be used to make precise measurements. The improvements of the remote sensor and digital camera promote the application of orthophoto in urban planning and GIS fields. However, there are still some problems with the traditional orthophoto, especially in geometric. The geometric problems of orthophoto mainly include the building tilt, the double mapping of building(ghosting) and the deficiency of space objects etc. The accuracy and the completeness of the orthophoto are influenced because of not considering these problems in the generation of DOM. As a result, when the vector maps of holographic and the orthophoto are superimposed on each other, the will be defects, such as the buildings are crossed by vector lines or contours of buildings and roads and dislocation to buildings and roads itself. It effects the analysis function of as the base map. The key reason of these problems is the occlusion between buildings or space objects.For the problems of the traditional orthophoto, true orthophoto as a new product of image map are emerges as times goes by. True orthophoto keeps vertical projection of the landscape, color uniform of image and makes full use of the function of images by eliminating projection distortion of terrain and spatial objects. Moreover it solves the occlusion problems that ground information is occulted by the shadow of orthophoto and tall buildings in large-scale urban areas.The occlusion detection is the most significant technology in the true orthophoto generating. The difference of true orthophoto generation depends on the different occlusion detection methods, to which many scholars pay their attention. So far it is the digital surface model and digital buildings model that are mainly used to detect the occlusion. In this paper, we research the generation of true orthophoto in two aspects combined the characteristics of data source with the existing algorithms. The main contents of the paper are introduced as follows:(1) The true orthophoto generation based on 3D model of digital cityFirst, multi-angle and multi-view images is preprocessed and block adjustment is implemented so that the coordinates of tie points of multi-view images can be obtained rapidly and accurately, then the 3D information can be acquired regular and irregular building geometry model are extracted, and the reconstruction of three-dimensional city modeling are completed. Second, to solve the geomatic problems of geometrical rectification, occlusion detection and occlusion compensation are considered as main research contents and true orthophoto is produced via geometrical rectification, and analyze and research the generation of true orthoimages, then the color consistency processing and image mosaic is carried out. Finally, analyze the factors that affect the accuracy of geometrical rectification according to the results of the experiments.(2) The true orthophoto generation based on DSMThe Z-buffer approach that has been proved to be efficient in detecting and dealing with hidden and occlude areas on the aerial images, so it is adopted in this paper. Based on this, we detect the occlusion areas through existed DSM data, and sign occlude areas for texture compensation. Finally, implement the color consistency procedure and image mosaic. At this point, the generation of accurate true orthophoto is completed.The information of occlude area, is compensated by using the visible region from adjacent images and selecting the optimal pixels to fill the missing region information. And finally, execute the color consistency and image mosaic procedure.