A Research Of Strategies For Template Registration And Decomposition Of Large Template Library For Mobile Augmented Reality App In Outside Architectural Environments

With mobile terminal processing capacity, bandwidth promotion, multimedia display performance enhancement and variety sensing module integration, augmented reality technology is widely applied to the mobile Internet network, intelligent terminal implements the augmented reality application and system miniaturization and more portable. Augmented reality technology, the mobile Internet technology, combined closely with the image recognition technology, increasingly mature technical scheme, will greatly promote the application of augmented reality application technology into the daily life. In recent years, Facebook, Google, Microsoft’s big tech giants have invested heavily, into the next generation computing platform, and this year the first year of assessment for augmented reality and virtual reality application, increased market reflect a large number of visual and virtual reality devices such as all kinds of glasses products, app store also appears to enhance traditional application interaction upgrade of all kinds of applications. The development of augmented reality application present from laboratory to life application, from large heavy equipment to diverse light mobile terminals, from indoor compactness spatial development to a wide range of outdoor environment. However, in view of the complex outdoor environment, especially the mobile augmented reality application of complex construction environment, only for simple individual object template recognition, object by looking at the picture of the entire area to identify objects, cannot meet the outdoor applications because of orientation, Angle and the recognition of local observation of buildings; In addition, to meet a wide range of outdoor applications, will identify the volume of a template library for linear growth in the number of recognition of the construction object, limited by the operation of the mobile equipment storage size and computing capacity, is too big, too complex a single template which will cause the mobile enhance recognition application in the process of identifying such problems as low efficiency and not smooth operation. In view of the above problem, this paper studies the outdoor registered model of complex building strategies, and on the basis of this research is based on the geographical position classification method template library depots, the main content of this paper including:(1) Campus outdoor augmented reality registration strategy based on typical building environment researchStudy typical complex building community environment, campus outdoor outdoor architecture based on template registration registration policy all party identification and application. Acquire the object of building ground plan, building on the graphics contour simplify rules, according to the shape of the graphics of the simplified, establish standards and complex building model classification. According to the characters of the specific shape of the model, the standard set to four types of basic graphics, such as:rectangular, "L" shape, "back" glyph and circular features, set up the corner position to solve the minimum circumscribed circle, use isometric segmentation method, calculating plane comprehensive building template registration point location. For complex combinations, first will be the combination of the subdivision standard graphics, establish rectangle, font and "back", "L" shape and circular of basic type of combination, then apply the method of calculating the standard type, calculation of architectural plane all-around template registration point location, the final registration point of the standard model combination merger, consolidation, in the process of aggregation of the same or similar registered point effectively, on the premise of guarantee adequate registration point, reduce redundant points registration.(2) Research based on the geographical location of template library depotsOn campus or broader augmented reality applications, such as city outdoor target identification library number registered with the size of the object into linear growth, will give the storage capacity and processing capacity of mobile augmented reality bring great challenge. For roaming through augmented reality applications, it explores application process has the characteristics of large scale of close to the ground, as a result, application of the location of the probe with the position of the obvious characteristics. In this paper, based on the geographical position classification method of template library depots in complex building model template registration point as the research object, the introduction of convex hull, for each object registration point to establish a convex hull, and on the convex hull of the area of the polygon object, the application of K-D tree, establish index K-D tree, to ensure that the number of registered point in each index node balanced, adapt and adopt the processing capacity of mobile terminal, based on K-D depots of the tree method, realize the recognition of the entire scene adaptive and efficient depots.Based on qualcomm augmented reality development kit, the typical application of campus building structure, build the test system, based on the standard and complex teach school registration and points to the geographical position has obvious partition registered template library depots to experiment, through the flat environment, such as mobile phones to actual validation, achieved good effect.