Estimasi jarak dan kecepatan pada pemain sepakbola berguna sebagai parameter dalam sports analytics. Dengan mengukur jarak (total distance covered) dan kecepatan, dapat diketahui performa yang dapat digunakan pada suatu tim. Penelitian-penelitian sebelumnya telah menunjukkan pentingnya mengetahui kedua parameter ini. Dalam ranah computer vision, tracking dapat digunakan untuk melakukan estimasi kecepatan dan jarak (total distance covered). Akan tetapi, hingga sekarang, belum ada penelitian yang membahas terkait performa estimasi kecepatan dan jarak dengan bantuan computer vision. Untuk mengisi gap penelitian, penelitian ini dirancang untuk menganalisis performa kombinasi algoritma object detection dan algoritma tracking untuk mengestimasi nilai kecepatan dan jarak (total distance covered). Penelitian ini menggunakan algoritma object detection YOLOv10, YOLO11, dan YOLO12 dengan ukuran nano dan small sebagai algoritma terkini untuk aplikasi real-time. Selain itu, digunakan algoritma SORT, DeepSORT, dan ByteTrack sebagai algoritma tracking dengan performa baik. Hasilnya, kombinasi model YOLO12s dan algoritma SORT menghasilkan nilai estimasi jarak (total distance covered) paling baik dengan MAE=3,438. Kombinasi model YOLO11n dan algoritma SORT menghasilkan nilai estimasi kecepatan dengan performa terbaik, yaitu MAE=0,966, RMSE=1,508, dan correlation coefficient=0.434. Hasil penelitian menunjukkan kombinasi model object detection YOLO dan algoritma SORT dapat menjadi salah satu pilihan dalam proses estimasi kecepatan dan jarak (total distance covered) pemain sepakbola.

---

The ability to estimate total distance covered and speed of soccer players is crucial. In sports analytics, these two parameters are essential for continuously monitoring both team or individual. By quantifying total distance covered and velocity, the performance statistics can be obtained. Previous studies suggest estimating total distance covered and speed is at most importance. In computer vision, tracking can be utilized to estimate the velocity and total distance covered. Nonetheless, up until now, there is no study that discussed the estimation of speed and total distance covered using computer vision. As such, to fill the literature gap and to investigate this unexplored topic, this research is performed to analyze the performance of combination of object tracking and object detection algorithm for estimating total distance covered and speed of soccer players. This research employs several object detection algorithms: YOLOv10, YOLO11, and YOLO12 with size nano and small as the recent developed algorithm real-time purpose. This research also employs several high-performing trackers: SORT, DeepSORT, and ByteTrack. Experimental results suggests that the combination of YOLO12s and SORT algorithm produce the highest performance in terms of total distance covered estimation, that is MAE=3.438. Also, the combination of YOLO11n and SORT algorithm achieves the highest performance in terms of speed estimation, with MAE=0.966, RMSE=1.508, and correlation coefficient=0.434. R. This research provides experimental basis as the combination of YOLO object detection model combined with SORT algorithm can be a choice for estimating speed and total distance covered of soccer player.