Кваліфікаційні роботи здобувачів вищої освіти кафедри авіаційних комп'ютерно-інтегрованих комплексів (НОВА)
Permanent URI for this collectionhttp://er.nau.edu.ua/handle/NAU/58725
Browse
Browsing Кваліфікаційні роботи здобувачів вищої освіти кафедри авіаційних комп'ютерно-інтегрованих комплексів (НОВА) by Subject "3D сцена"
Now showing 1 - 2 of 2
- Results Per Page
- Sort Options
Item Image depth detection system(National Aviation University, 2022-11-23) Sych, Oleksiy; Сич, Олексій СергійовичIn today's world, there is often a question about creating a model to solve a certain problem in such a way that it performs its intended task properly and does not have a large cost. This is what almost every developer of the project wants at the production stage. Thus, the work consists in improving and acquiring better accuracy of the image depth detection system. For this, was modified and improved, namely, the main design of the model was changed and the quality of the image was improved, thanks to various methods of image filtering. Unlike the previous model, this project investigates the effect and quality of the 3D scene construction in the image, not the streaming video, under different weather conditions and at different observation points, which makes it possible to feel in more detail the impact of various phenomena on the model itself during operation and improve accuracy due to considering a single pair of images rather than a stream of large numbers at a specific frequency. The design consists of two cameras, which were selected from the principle of price-quality, and a box to fix and protect the model itself, thus forming protection from the environment in various conditions of use. The design is connected to a computer that performs the software part, which consists in creating a stereo pair – artificial adjustment of cameras, image analysis at the initial stage and after filtering, which as a result gives an opportunity to see the difference in the accuracy of constructing a 3D image, which can be used for various goals, for example to find out the size or distance to the target object.Item Image depth estimation system by streaming video(National Aviation University, 2021-06-16) Сич, Олексій Сергійович; Sych, O. S.Today, the tasks of computer vision are becoming very relevant, more and more people are automating work in production due to some kind of software processes and machine devices, which can make job easier or more accurate. Based on this, it was decided to consider in detail the problem of stereo vision without using neural networks, or other more complex methods, since their use required costly methods of training, setting and controlling parameters. The main task was to create a mechanism taking into account the price and quality, due to the fact that there is no cheap analogue on the internet market, which was suitable for the task of simple recognition of 3D scenes and made it possible to analyze the environment in which it is located, namely, to find out at what distance objects are located, what is their size, and so on. In the course of the work, the method of using two web cameras was chosen, which were configured and calibrated for the task of stereo vision. The conditions of projective geometry and the relationship between the two cameras are also considered, since without this, the operation of the main algorithm of the work could not be successful at all. An algorithm and program have been created for the device to operate in streaming mode, which allows directly know the exact characteristics in LIVE video mode.