Unfortunately few places are as heavily polled as the US, so I've not been able to find a lot of data for Japanese opinion of article 9, but I have found a few. The results I have found appear to be mixed.
From a 2014 article (emphasis mine):
This policy is not intended to do away with postwar Japanese pacifism.
The word “pacifism” is not included in the English expression “policy
of Proactive Contribution to Peace,” but Abe’s expression
sekkyoku-tekiheiwa-shugi in Japanese implies positive pacifism.
Sekkyoku-teki literally translated means “proactive” and heiwa-sugi
literally translated means “pacifism.”
Japanese public opinion has, however, not been very supportive of the
decision. An urgent opinion poll conducted on July 2 and 3 by the
Yomiuri shimbun, a national daily newspaper with Japan’s largest
circulation (9.25 million in July 2014), said that the approval rating
for the Abe administration was 48%, down 9 percentage points from 57%
in the previous opinion poll in May. The disapproval rating was 40%
(31% in May’s survey). In the Yomiuri shimbun’s opinion poll, the Abe
administration’s approval rating fell below 50% for the first time
since the administration took office in December 2012. In the poll,
51% did not support Japan’s limited exercise of its right of
collective self-defense, against only 36% who did support it. Opinion
polls in other media outlets showed similar results. A poll conducted
on July 11 to 13 by NHK (Nippon Hoso Kyokai [Japan Broadcasting
Corporation]), Japan’s semi-national public broadcasting organization,
said that 38% supported the cabinet decision to approve the exercise
of the right of collective self-defense, while 56% did not. The Abe
cabinet was supported by 47% of respondents, down 5 percentage points
from a poll in June. For NHK polls as well, this was the first time
the approval rating for the Abe administration had fallen below 50%.
(In the poll, 38% of respondents did not support the administration,
up 6 percentage points from the previous month.)
...
An opinion poll conducted by the Asahi shimbun on May 24 and 25
showed that only 29% of respondents supported the change in
constitutional interpretation to approve the exercise of the right of
collective self-defense, which the prime minister sought. Exercising
the right of collective self-defense was opposed by 55%. The results
appear to suggest that the majority of voters in Japan are strongly
opposed to Japan’s exercise of collective self-defense.
...
However, the result of a poll jointly conducted by the Sankei shimbun and the Fuji News Network (FNN) on May 17 and 18 gives a completely different impression. In this poll, 69.9% of respondents supported Japan’s exercise of the right of collective self-defense (10.5% of respondents answered, “The right should be able to be exercised in full scale,” and 59.4% answered, “The right should be able to be exercised to the minimum extent necessary”). The exercise of the right was opposed by 28.1%. A poll conducted by the Yomiuri shimbun on May 30 through June 1 shows similar results. Exercising the right of collective self-defense was supported by 71% of respondents (11% of respondents answered, “The right should be able to be exercised in full scale,” and 60% answered, “The right should be able to be exercised to the minimum extent necessary”) and was opposed by 24%.
In a monthly poll conducted by NHK on June 6 through 9, 26% of respondents answered, “Japan should be able to exercise its right of collective self-defense,” and the same percentage of respondents answered, “Japan should not be able to exercise its right of collective self-defense.” The largest percentage of respondents, 46%, answered, “I cannot say either way.” As for Abe’s approach to enabling Japan to exercise its right of collective self-defense by changing the government’s traditional constitutional interpretation instead of revising the constitution, 22% of respondents supported the approach, 33% were opposed, and 40% answered, “I cannot decide between the two.”
Note that this article goes into some explanation as to why the results seem to vary so much; I recommend reading the full article.
From a 2016 article (emphasis mine):
The Yomiuri Shimbun poll in March revealed that 49 percent of its respondents supported revision while 50 percent opposed it. The Asahi Shimbun reported in May that 37 percent of its respondents opted for revision and 55 percent were against it. Finally, the Nikkei reported also in May that 40 percent of its respondents were pro-revision while 50 percent were not.
This article also has some useful graphs of opinion polls over time given by the major newspapers:
