Rotating the image and stretching the pixels you get a pattern (pic 1). Adding the photo finish behind it gives a perfect match (pic 2). This is the exact spot which the finish camera captured.

Can we use the video camera on the finish to estimate the distance to the line? Roughly yes, as the finish line centers well in the image.

Matching a line with the 'o' and the dark spot gives the next result:

As the finish line didn't fully cover the road, we drew the finish line trough to the barriers and use it to roughly estimate the gap between the finish line and the photo finish

The finish is 96 pixels modelled at the barrier, the gap between the two lines is 71 pixels.

71 / 96 * 100 = 74% of the line width is the offset from the photo finish spot.

As defined in the UCI regulations, the width of the line should be 34cm.

75% of ~34cm is an offset of ~25cm.

Finally, using the same percentage again on the other side of the road to create a point to which you can draw a straight line, you can get a good estimation of the photo finish line coloured green:

.Nobuo Ozaki has also completely modelled the finish with the shot by using the width of the finish line (34cm + 4cm + 34cm). The results are almost identical: 240mm

La Flamme Rouge Nobuo Ozaki Pidcock won...that is clear from all of this.

@allemagwzime La Flamme Rouge Nobuo Ozaki Most straight on photos show Pick in front...so there's that

anenglishmaninspain @allemagwzime La Flamme Rouge Nobuo Ozaki Van Aerts wheel is up in those. You can't judge it with one wheel up and one down