Background: The use of digital cameras set-up and large calibration volumes for three dimensional (3D) analyses in controlled indoor or laboratory settings is common, mainly considering cyclic sports. So, for the best accuracy in those 3D scenarios, it is crucial to test the best combination points in a large and static calibration volume. Aim: We aimed to verify the Root Mean Square Error (RMSE) in a large and static calibration volume using 300 randomized coordinates for horizontal (x), vertical (y), and lateral (z) axes calibration points. Methods: A stainless steel calibration volume (6000 x 2500 x 2000 mm) with 108 markers attached with black tape at each 500 mm positioned on a basketball court and recorded with four digital cameras (50 Hz; Go Pro Hero 7 Black). The Matlab software (MathWorks®) was used to plot 33 matrices, 11 for each axis containing twenty to thirty elements with random coordinates of the calibration volume. Results: From 20 to 30 points the range of RMSE was 25.07 to 26.09 mm (x), 14.38 to 14.32 mm (y), 14.22 to 14.36 mm (z) and 32.53 to 32.77 mm for the three axis resultant. The minimum and maximal RMSE for x, y and z axis were 9.05 to 44.09 mm, 9.33 to 18.10 mm, and 9.00 and 18.09 mm. The minimum and maximal total RMSE was 16 to 51.37 mm. Conclusions: As preliminary findings, the algorithm has shown potential to return sensible RMSE for the y and z axis in the aerial. Further tests are needed to approach more coherent RMSE values for the x axis, and consequently, the total error in aerial and underwater environment. This study was supported by grant: Amazonas Research Support Foundation (FAPEAM), Infrastructure Program for Young Researchers - First Project Program PPP nº/004/2017 (grant number 062.01554/2018).
© Copyright 2026 XVth International Symposium on Biomechanics and Medicine in Swimming Proceedings. Alle Rechte vorbehalten.