By "kill" do you mean that 1) on one day, someone was just walking down the sidewalk, happy-go-lucky, prime of life, sunny day, and the next day just died?
Or 2) It's estimated that a person who hasn't done too much excercise or is a smoker might have lived to 73 and actually died at 72.5 years?
I'm guessing it's somewhere between the two, but I don't see the details needed.
I did some digging, and for lung cancer, it seems the average age of diagnosis is 71. Considering that life expectancy is about 78 years in the US, and assuming 1 year average between diagnostic and death (the 5 year survival rate of lung cancer is pretty low, so lets err on the side of overstating the impact), that gives an estimated 6 years lost per victim, or the equivalent of 215 full lifetimes lost per year.
Not sure what the expected Fukushima cancer death toll will be. It looks like the median age is about 45 years in Japan. Guess about 10 years between exposure to low-lying radiation and cancer. Say it's as deadly as lung cancer, and on average kills the person 1 year later. Life expectancy in Japan is 83. That's 83 - 56 = 27 years lost per victim. So if Fukushima, on average, would cause about 510 additional (Japanese) deaths each year due to cancer, it would be the same amount of years lost as is due to lung cancer via coal power plant pollution in the US. (Presuming no other nasty nuclear accidents in Japan at the time.) Admittedly, we're comparing apples and oranges, somewhat, since the population of the US is 2.6x that of Japan. So really, 196 additional cancer deaths each year would be needed to put the Japanese nuclear industry (including Fukushima) on par with the US.
First google search for estimated Fukushima deaths from cancer puts the most likely number at 130 total. (Not per-year.)
All of this post is admittedly a back-of-the-napkin calculation with several guesstimates. There's also some flaws in the methodology. I'm also comparing a smaller country with a higher population density but smaller total population to a much larger country with a lower population density but higher total population. But it seems that for the risk of cancer deaths, the Japanese nuclear industry is far safer, including the meltdown, than the US coal power industry, unless one of my guestimates was wildly out of line. Even if I'm off by a factor of 10 in underestimating the risk of Fukushima, or overestimated the risk in coal, Fukushima-catastrophes would have to strike about once a decade in Japan to have a similar amount of years lost in proportion to the total population.
There's a few other factors involved as well. I ignored the 600 evacuation deaths from Fukushima. That's a one time event, but it does up the amount of deaths from Fukushima significantly. On the other hand, I haven't considered the majority of estimated deaths from coal power is not due to lung cancer. Assuming that the non-cancer deaths from coal are similar to the cancer deaths (6 years lost per victim), and figure that the evacuation deaths are evenly spread out among the Japanese population (39 years lost per victim). We'll go with about 10,000 dead from coal power that isn't lung cancer (going with the later study that shows a lower death toll), and that gives us 60,000 years of human life lost each year. While Fukushima's 600 evacuation deaths are at around 24,000 years of human life lost. But remember, US has 2.6 times the population, so Fukushima so it actually works out as slightly more years (60,840 years vs 60,000 years) of human life if the populations are equalized.
There's also the environmental and economic effects. Fukushima took out a few hundred square miles of land due to fallout. That's going to have a real cost. OTOH, we could estimate the land lost to global warming, and figure out carbon power plants in the US's share of the pie, and that's going to have a real economic cost as well. No comparison will be perfect. And we could redo all of the above calculations with more accurate figures, but I strongly suspect the result will be in the same ballpark.
I'm rather disturbed by the result I have. Unless I've made a horrible mistake that throws off the result of the calculations, in terms of years of life lost when compared to the total population size, coal in the US is more or less on on par with a Fukushima-event happening every year in Japan.