ALGORITHMS OF FINDING OUT OBJECTS AND ESTIMATION OF PARAMETERS OF THEIR MOTION ARE IN SYSTEMS OF TECHNICAL SIGHT
Most contemporary machine vision systems process video data contained in the recorder that is installed on a mobile platform. These designs are oriented to use the single-board computers Raspberry Pi and the Python programming language combined with OpenCV image processing library. The Python and Op...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
National Aerospace University «Kharkiv Aviation Institute»
2019-03-01
|
| Series: | Радіоелектронні і комп'ютерні системи |
| Subjects: | |
| Online Access: | http://nti.khai.edu/ojs/index.php/reks/article/view/623 |
| Summary: | Most contemporary machine vision systems process video data contained in the recorder that is installed on a mobile platform. These designs are oriented to use the single-board computers Raspberry Pi and the Python programming language combined with OpenCV image processing library. The Python and OpenCV couple allows real-time video data processing while implementing modern algorithms. When doing this, there is a need to provide for high accuracy and fast response of information parameters measurements (machine vision systems should run in real time). The problem of moving objects reliable detection and appropriate parameters evaluation is especially topical in presence of interferences those resulted from scene lighting variations. Such a complex of troublesome factors were taken as a base of the proposed development, the results of which can be used for solving various control problems in robotics engineering as well as accurate positioning of unmanned aerial vehicles.
Algorithms are offered for moving objects detection and following motion parameters determination that are based on video monitoring (video camera surveillance) outcomes is proposed. A conception on synthesizing algorithms for moving objects detection and following appropriate parameters determination, which utilize video observations providing for image binarization on the base of HSV brightness-color components analysis is proposed; the binarization implements the image clipping on a brightness threshold and the following binary image moments detection that allows precise localizing the object within the frame A simple and accurate technique of the object trajectory parameters evaluation (distance to object, and angular and velocity parameters) has been proposed. Measures to raise thresholds of steady running of the algorithm realized by using an interactive interface are provided. Measures to raise thresholds of steady running of the algorithm involving an interactive interface are provided. Investigation outcomes could make use for designing various machine vision systems. |
|---|---|
| ISSN: | 1814-4225 2663-2012 |