Numerical Model - Methodology

Numerical Model – Methodology

In order to carry out my Numerical Analysis, I had to use 2D Video Analysis in order to assess the angles and distances which were relevant to my athlete’s skill, in this case, a netball shot. The main software I used in order to carry out this analysis was Kinovea which allows me to measure angles and distances of joints, release heights and the athlete’s height. However, to be able to use Kinovea in this way, a number of steps must be taken first in order to correctly prepare the video footage for importing into the software to be analysed. Other precautions must be taken to prevent errors which could affect the reliability of my data analysis.

The stages of 2D Video Analysis can be displayed in the following flow chart:

Once the video has been imported into Kinovea, I merged the scale video and the ‘skill’ video on top of one another so that I could create a 1m scale in Kinovea. This allows me to create lines and angles on the video and the programme will provide me with a ‘real life’ measurement.

It’s also important that the video recording takes place in front of a plain background with no additional lighting. This is so that the anatomical markers are clearly visible once the video has been imported into Kinovea. This makes measuring joint angles and body distances much clearer and prevents the occurrence of errors.

Scaling is an important step when performing 2D Video Analysis as it prevents perspective and parallax errors occurring which could hinder the reliability of my results. Without scaling, displacements would be measured in ‘pixels’ by the programme and not measured in M and Cm which we require. This also makes the comparison of subjects much easier if they are measured by the same unit.

Once the scale has been set, it’s important not to move the frame of capture away from this point as we risk perspective error. Perspective error results in subjects appearing larger or smaller than they actually are. Using a scaling factor means that we can obtain realistic measurements (i.e 1.2m) rather than the 2-3 cm that it may appear as on the screen. If we move the frame of capture away from the location in which the scale was created, when the video is imported into Kinovea the 1m scale will not actually be 1m in ‘real life’. As a result, our analysis cannot be considered valid or reliable and its makes the feedback we can provide in-effective.