CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
Usage of computer vision for monument recognition through mobile application “electronic guide”
 |
Read | Kuznetsov M. A., Vorobyov A. S. Usage of computer vision for monument recognition through mobile application “electronic guide” // Caspian journal : management and high technologies. — 2018. — №2. — pp. 25-36. |
Kuznetsov M. A. - Cand. Sci. (Engineering), Volgograd State Technical University, 28 Lenin Ave., Volgograd, 400005, Russian Federation, mara122@mail.ru
Vorobyov A. S. - post-graduate student, Volgograd State Technical University, 28 Lenin Ave., Volgograd, 400005, Russian Federation, a.vorobyov.a@yandex.ru
The article offers a technique for developing a descriptive set of monuments out of a collection of their photographs in order to organize a database for automatic recognition. Described technique is used in the form of mobile application “Electronic Guide” based on AR platform Vuforia. This platform is used for database creation and storage, identification and tracking objects of urban infrastructure and for implementation of interactive user interface with augmented reality. Object recognition mechanism is based on computer vision algorithms. The mobile application was implemented in Unity3D development environment and tested on Android platform. The article describes the results of experiments with the recognition of Volgograd monuments. On the basis of experiment results, it was concluded that the described approach is efficient for solving the problem of determining the objects of the urban infrastructure. It has many advantages including high recognition accuracy. However, weather and lighting conditions can influence on application work quality. In addition, the formation of marker base is laborious. Furthermore, it is necessary to separate recognition errors into the 1-st and 2-nd kinds.
Key words: распознавание образов, алгоритм распознавания, ошибка распознавания, погодные условия, условия освещения, дополненная реальность, маркер, Vuforia, Unity3D, информационно-телекоммуникационные технологии, база данных, object recognition, recognition algorit