Using vector building maps to aid in generating seams for low-attitude aerial orthoimage mosaicking: Advantages in avoiding the crossing of buildings

Abstract

Abstract:A novel seam detection approach based on vector building maps is presented for lowattitude aerial orthoimage mosaicking. The approach tracks the centerlines between vector buildings to generate the candidate seams that avoid crossing buildings existing in maps. The candidate seams are then refined by considering their surrounding pixels to minimize the visual transition between the images to be mosaicked. After the refinement of the candidate seams, the final seams further bypass most of the buildings that are not updated into vector maps. Finally, three groups of aerial imagery from different urban densities are employed to test the proposed approach. The experimental results illustrate the advantages of the proposed approach in avoiding the crossing of buildings. The computational efficiency of the proposed approach is also significantly higher than that of Dijkstra's algorithm.

Automatic Seamline Determination for Urban Image Mosaicking Based on Road Probability Map from the D-LinkNet Neural Network

“…The seamline In many related studies, it is difficult to find a general method for automated quantitative assessment of the seamline quality. Therefore, similar to the evaluation method in other relevant studies [2,6,7,25], the quantitative index applied in the proposed method is the number of times that seamlines cross obvious objects. The seamlines which have a smaller number of times are considered ideal seamlines.…”

“…Therefore, more training samples should be made to better train the D-LinkNet neural network. (2) The proposed algorithm can be applied to the seamline network determination framework [3,46,47] to construct a single seamless composite image automatically.…”

“…However, with the development of technology, the orthoimage spatial resolution becomes higher, and the coverage area of an individual orthoimage is typically very small Thus, image mosaicking is a necessary process of constructing multiple images into a large-scale and single seamless composite image. This process has been applied in a wide variety of applications such as environmental monitoring, agricultural monitoring, and disaster management [2,3]. Orthoimages are typically orthorectified by the Digital Terrain Model (DTM) of the same geographical area.…”

“…Chen et al [17] first used the Digital Surface Model (DSM) and Digital Terrain Model (DTM) to derive an Orthoimage Elevation Synchronous Model (OESM) that accurately reflected the pixel of each digital orthophoto image, and then obtained the final optimized seamline using the Dijkstra algorithm. Wang et al [2] used vector building maps to determine the seamlines, which guaranteed the seamlines avoiding the crossing of buildings as much as possible. Different from the method of tracking vector roads, the seamlines determined by this method went along the middle line between buildings in order to avoid crossing the obvious objects, especially for high-rise buildings.Using the normalized difference vegetation index (NDVI) and morphological building index (MBI), Pan et al [18] introduced ground object classification into the seamline optimization method.…”

Automatic Seam-Line Detection in UAV Remote Sensing Image Mosaicking by Use of Graph Cuts

“…It uses graph cuts to obtain the best mosaic texture on both sides of the seam-line to obtain the optimal seam-line, which can solve the shortcomings of dynamic programming in the image mosaicking (such as no need to specify additional energy end points, etc.). In addition, compared with the annealing algorithm and ant colony algorithm, it has the advantage of fast calculation [20][21][22][23][24][25][26]. In these optimal seam-line search methods based on graph cuts, they can be divided into a color domain-based method, a gradient-based method, and a color domain and gradient domain-based method according to the difference of the energy function smoothing term.…”