Use the H2 database in junit-benchmarks to catch performance regressions

It is possible to use the embedded H2 database to store previous runtimes and so make beautiful history charts as well as measurements that make tests fail when the runtime is substantially worse than before.

This needs to be modeled after the [unit tests of junit-benchmarks](https://github.com/scijava/junit-benchmarks/blob/master/src/test/java/com/carrotsearch/junitbenchmarks/h2/TestH2HistoryChart.java) because the [documentation is insufficient](http://labs.carrotsearch.com/junit-benchmarks.html).

Start by reading junit-benchmarks' TestH2HistoryChart.java, which the issue identifies as the model, then inspect the imagej-ops benchmark tests for the integration point. Done means previous runtimes are stored in H2, history charts and measurements are available, and substantially slower runtimes cause tests to fail.

Written by the indexing model from the issue text.