There are 3 initial situations that all require different approaches, depending on the involvement of  illustrations created from scratch, raster files, or legacy vector illustrations without object assignment.

Illustrations created from scratch
The easiest way of getting an intelligent illustration is to draw it from scratch. Each newly drawn graphic should be created with graphical objects, even if their actual use can only be determined at a later point in time. When naming objects, one could use the spare parts number of the part, for example.

Raster files
Manually drawn illustrations must be scanned in order to use them electronically. These scans, or generated raster files, do not contain hotspots or objects. However, if they are saved in a hybrid file, you can manually create graphical objects that can later be treated the same way as vector objects. 

If there exists a sufficient number of raster files with appropriate hotspot overlays in an existing catalog, the automatic conversion to graphical objects is possible. For this purpose, a special plugin can be developed, for example for IsoDraw, which allows overlay data to be converted to graphical objects and,  together with the raster graphic, saved in a hybrid (e.g. IsoDraw) file.

Legacy illustrations without assigned objects
The major share in illustration archives consists of vector illustrations without assigned graphical objects. Data that was drawn with an illustration program like IsoDraw are counted among these. In these cases there are two ways of assigning graphical objects.

In the first case, the object attributes are assigned manually to the already drawn groups. This procedure is conducted once and does not have to be repeated during revisions. Newly added parts in the drawings are created as graphical objects from the beginning.

Secondly, there is the possibility of automatically converting text elements to graphical objects. In this case, a specially developed plugin for the illustration software, for example for IsoDraw, searches the whole illustration for text elements and converts these into graphical objects according to certain, pre-defined criteria. Recognition of position numbers serves as an example.