Relevance. The urgency of developing an automated method for rapidly and reliably determinining shooting results during sports competitions arises from the need to significantly reduce the time spent by sports referees on the results determination commission’s (RDC’s) process of analysing shot paper targets. This will help eliminate subjective errors and improve the overall quality and transparency of refereeing. Objective. To develop and implement digital technologies, particularly computer vision algorithms based on You Only Look Once (YOLO) neural networks, in order to automate the time-consuming and repetitive task of assessing the quality of shots and determining overall results at paper targets during large-scale sports shooting competitions. Methods. To achieve this goal, we used the following methods: theoretical analysis of scientific literature and regulatory framework on the stated issues; modeling of the target processing process; computer vision techniques (object detection in real-time based on the YOLOv11 neural network); geometric algorithmization for calculating the shooting results; and an experiment to test and verify the accuracy of the developed software in mobile device applications. Results. An original automated method for evaluating the results of sports shooting has been developed. The method is based on a novel algorithm that integrates a mobile platform, the Assistant RDC Android application, with the YOLOv11 model for detecting bullet holes of various calibers and a geometric model for unambiguous classification of the holes based on their size. This approach eliminates subjective errors in scoring and significantly improves the speed of evaluation of participants’ scores. Conclusion. The proposed digital transformation of the shooting sports refereeing system addresses significant drawbacks of the traditional system, which relies on visual inspection and manual calculation of target scores. The new system ensures consistent and reproducible accuracy, comparable to that of a skilled sports referee, while eliminating human error.
Тараканов et al. (Thu,) studied this question.