In the world of procedural programming, a pipeline is a subsystem of a program that takes data from one "end" and does something with/to it, possibly resulting in output data at the other "end".
Many old school programmers like to think of their programs like they were a kind of a factory, where stuff comes in from the other end (either interactively or using a batch mode of some sort), is put through several more or less independent processing machines after which the final output is stored somewhere (or shipped out, where applicable). If you were to draw a data flow diagram for such a program, it would look much like a complex pipe system which handles data instead of water.