Image processing, be it for geographic applications or not, generally requires the execution of a sequence of transformations (that is, steps) in order to obtain the desired final result. In these sequences, the output of one step is the input of the next one. In computing, this is called processing pipeline.

This type of data manipulation is very versatile, because you have a range of functions or steps than can be arranged into numerous combinations to produce a wide range of results.

What we did so far, in this chapter's examples, was we opened an image from the disk, performed a given operation, and saved the results to another image on the disk. Then, in the next step, we opened the result from the previous one and so on.

Despite the steps are not yet connected, we can imagine the following image processing pipeline:

Saving intermediary steps to the disk is useful when we want to use the images from them, or in other situations, when the pipeline uses...