[Documentation] Improve pipeline documentation

[russelldj]

The pipeline documentation seems quite obtuse to me. Since this seems like a primary way end users would interact with KWIVER, it seems worthwhile to make it both more accessible and easier to read. For example, I'm still confused on the exact taxonomy of a process block and my read of this doc does little to address my questions. Perhaps starting with that and other basics, like how to use the pipeline runner and plugin_explorer, would be more useful. It seems like then many of these current topics could land in an Advanced section, or something analogous.

[russelldj]

I guess I was roughly looking for the process definition section, which is sort of buried in other stuff. It also doesn't explain how to use the type specifier, which seems to some times be prefaced with a different name than the algorithm type. E.g., an image_filter type would be specified as filter:type.