This contain a compilation of other videos from similar perspectives but I see no evidence of the same flashs:
http://www.youtube.com/watch?v=MJXfIOEQB8M
Unless some other evidence turns to support the existence of those particular flashes, I guess I am going leave be unexplained video anomalies.
The reason I do not think they are simple compression artifacts are:
(1) They are not "interference pattern" type effects, where the length scale of the image features is near that of the size of a pixel or the compression algorithm sampling length.
(2) The output of the compression algorithm should be the same up and down the length of the tower, so artifacts should not be vertically localized.
(3) Compression artifacts should not be in sharp contrast to surrounding pixels, in this case brighter, rather they should be near average contrast of the surrounding pixels.
But again, at this point I writing these particular flashes off to some kind of unknown video anomaly because I do not see them reproduced in the video compilation I linked above.