Update (Nov 17): upon a few more viewing, I have come to the conclusion that Minjinyu also made a turn of about 20-30 degree and at some point was not exactly facing left. I have updated my graph accordingly. Note also that I am not trying to prove that Minjinyu bears no responsibility at the collision. I believe both sides need to share the responsibility, since Mizuki also made a significant turn of some 240-270 degrees, and made its direction very unpredictable -- might have caused the seemingly erratic turn of Minjinyu. (But we really do not have solid evidence to say either way)
But before you go on I want you to first judge this picture, that if the boat has turned
Then you can find out the answer in this picture below (where the picture above is cropped from), taken from the end credit of the movie Suspect X (a great movie btw). I do not claim the situations are the same, but this just shows us how a partial picture is far from being a conclusive evidence. That is why I hope Japan would release all the 10 hours of videos it has taken.
In my previous post an anomymous commentator suggested that Minjinyu 5179 made a change of direction and hence ram into Mizuki. What he saw was the video taken from on board Mizuki, all it shows is the relative velocity. Fortunately, we are able to see another video taken by Hateruma during the exact same moment from a different angle. Here it is.
This is a screenshot taken by Japanprobe, from the video taken by Hateruma, observing the collision between Mizuki and 5178. You can see the trace behind Mizuki, when it just completed a 300 degree turn right in front of 5179. It is quite clear that it is Mizuki which had made a large angle turn just before the collision, towards the area blocking the path of 5179
A path with time scale based on the video, and the speed indicated by the intervals between the 10 second gaps is shown approximately here. Looking at the time scale you can also see that Mizuki has more or less parked (stationary speed) for the 30 seconds before the collision, while Minjinyu had actually slowed down a bit. You can view the video again to verify my chart.
1A) Mizuki speeding towards 5179 (0:47)
If you compare the location of Mizuki between 0:26 and 0:29, in less than 3 seconds it traveled the length of its own of 46 meters (type びざん型巡視船 (2代))(marked by the white wave). So its speed is about 50m/3s = 60km/hr, 4 times the speed of fishing boat 5170. So even adjusting for it slower speed right after the turn, it must be able to avoid the crash it wanted to.
If this is still not conclusive, in the Mizuki video we can see two other JCG staff were taking videos, they should release those videos as well as additional data points.
Now back to the video taken from Mizuki.
A second JCG ship can be seen on the right of 5170, trying to encircle it from left behind
The encircling tactic is not uncommon from JCG. This is an aerial photo from another act by JCG, probably taken in 1998 in the same area. This may be the plan of the JCG ships.
See also a different view in this NBR thread, there are some problem with the poster's observation though, e.g. he claimed there were 2 JCG ships, in fact there were at least 3, the Yonakuni, the Mizuki, and the Kateruma.