INTRODUCTION: During competition analysis, individual distance swam during different race segments have been scarcely measured in competition (Pai et al., 1984). The aim of the study was to check the accuracy and reliability of a system for competition analysis in swimming based on individual distance measurements, comparing it with the most commonly used scaling technique. METHODS: 128 swimmers participants in 100 meters events of all four strokes were filmed in the 50x25 meters pool. Two competition analyses were performed simultaneously, based 1) on a linear scale system and 2) on 2D-DLT photogrammetry. RMSE accuracy of the position and intra-control point`s distance was assessed. Repeated digitization of hand entry and head emersion was checked for consistency by mean of the coefficient of variation. Race segment times using both procedures were compared with Bland and Altman`s 95% limits of agreement. RESULTS: RMSE when reconstructing the 2D-DLT position of 32 control points was 0.050 meter, less than 0.5% of control space in x axis. The reconstruction of intra-control point`s distance showed RMSE=0.046 meter, less than 1.2% of the total distance. The repeated digitization of the freestyle turn`s total distance in each lane of the pool showed coefficients of variation less than 1%. Maximum systematic differences between 2DDLT and scaling technique occurred during freestyle and backstroke turn time (0.05 s). Maximum random error occurred always in breaststroke being 0.17 s during start time, 0.21 s during swim time and 0.23 s during turn time. DISCUSSION: 2D-DLT showed great accuracy values in the present study with errors similar to other studies utilizing the same technique (Challis, 1998). Two technical actions successfully tested during the digitization process allow measurements of any variable during competition. However, no other studies had previously applied DLT techniques to swimming competition analysis. Differences between 2D-DLT and scaling technique (no longer than a frame of standard-speed video (0.04 s.) could be acceptable for its practical use in competition analysis; however, precautions must be taken when measurements are made during underwater parts of the race.
© Copyright 2010 Biomechanics and Medicine in Swimming XI. Published by Norwegian School of Sport Sciences. All rights reserved.