Methods of creation and practical application of mask -maps of high-level terrain objects at orthotransformation of digital aerial photographs

Aim. To develop and accomplish an experimental testing of the method of creating a mask map of high-level terrain objects. Methodology. On the basis of the cross-correlation method, it is proposed to carry out the estimation of the similarity of orthophotos that have a mutual spatial overlap. The comparison of the left and right images takesplace in pixels, for pixels with the same spatial coordinates X and Y, and therefore, there is no need to organize the movement of the search bar and to search for the corresponding points as such. It also removes the limit on the size of the image – the standard. In addition, the relatively small cross-correlation sensitivity to the differences in the illumination of the scene is very important for the choice of the correlation coefficient as a measure for the comparison of images. Taking into consideration the perspective deformations of images of high-level terrain objects it is expected that the number of pixels with negative comparison is significantly higher for regions with images of such objects. The overall picture throughout the study is a map-mask of high-level terrain objects. Such a map can be formed with the help of geoinformation modeling of polygonal objects, which outline zones with a high compaction of pixels with a negative comparison result. Results. The considered method of creating a map-mask of high-level terrain objects provides a possibility of obtaining important information about the quality of the digital topographic surface model used for orthotransformation of aerial photographs. The revealed effect of compaction points with a negative result of the mutual comparison of orthogonal images by the cross-correlation method allows to identify and establish the spatial location of high-level terrain objects such as roofs of buildings, fences, power lines, crowns of trees, and shrubs. Practical meaning. An example of map-mask application of high-level terrain objects for orthoimage stitching is given.