| Title |
Krepšinio treniruotės metimų pataikymo automatinio atpažinimo tikslumo tyrimas / |
| Translation of Title |
Accuracy analysis of automatic counter of basketball practice shoots. |
| Authors |
Toliatas, Ernestas |
| Full Text |
|
| Pages |
60 |
| Keywords [eng] |
automatic recognition of attempted and made shoots ; automatic counting of attempted and made shoots ; automatic tracking of objects ; automatic analysis of basketball practice in the real time ; auto filter of colors |
| Abstract [eng] |
In this master thesis automatic recognition system of made shoots (ARSMS) was created which works in the real time or from video file. Also analysis of ARSMS accuracy and tracking methods was made. System was created using open source library OpenCV and third party software IP WEBCAM, which was needed to get video from telephone camera to OpenCV and C++ programing language. Because of that data transferring methods were analyzed using wi-fi and bluetooth networks. It was created object tracking technology which used recognition of colors and motion detection alongside with auto filter of colors. The system can track limited amount of people and basketball. It was analyzed performance of software when one camera was used from side and behind when only ball was traced and ball with people. Analogical experiment was made for case when two cameras were used. The last version of software tracks people and the ball with two cameras. Camera from behind tracks people and ball, camera from side tracks only ball. Attempted and made shoots is being counted for every person and statistical evaluation is made. Also software tracks if every person is keeping up with goals which were set up before practice. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2018 |
