Normalization is a design consideration, and is based on organizing YOUR data for minimal redundancy and better performance. Because it is data dependent, there is no automatic process, you have to apply knowledge and rules. Certainly you can use the table metadata and the actual data to find possible areas of un-normalized data, for example nullable fields, null data, repeating data etc.
To add to @Kev Riley's answer, what's even better is that what may be normalized for one table may not be for another. A great example is telephone numbers: supposing that you have just American phone numbers with no extensions, they're in the format (xxx) xxx-xxxx. Is this normalized? It depends... Are you going to analyze area codes and try to derive geographical information? If so, then no, this is not normalized. But if you're just printing the telephone number on an invoice and never doing anything else with it, then you could make an argument that it is normalized. Unfortunately, aside from some really obvious cases (like a table with Field1, Field2, Field3, Field4), normalization is a process, not a formula.
To expand slightly on Kev's excellent answer, there is no way to automate normalization starting with a database since proper normalization requires an understanding of dependencies within the data. Determining the dependcies requieres an understanding of the data and the relationship between the fields, this part cannot be automated. Now, if you can fully specify all functional dependencies within your data then it does become possible to automate the normalization at that point. However, in almost all real world cases precisely specifying the functional dependencies is the hardest part of normalization, and putting it into a form that can be used by the automated tools can be painstaking even when you know what allt he dependencies are. So, it is technically possible to automate it if you are willing and able to specify every functional dependency, but this really only comes up in academic/theoretical situations. As a practical matter, it is generally faster, simpler, and just as effective to do it by hand.