Is there any algorithm available to automatically classify the builtup area in Arcpy?

[FarhanMustafa]

I am working on a project in which I have to classify high resolution images to extract the built-up areas of 50 cities and then to calculate the percentage of the area. Is there any algorithm existing in Arcpy which helps to classify the built-up area automatically? I mean if I give the raster image as input and the output should meet the requirement. 

[XanderBakker]

The tools that you have at your disposal (with a Spatial Analyst license) are in the Multivariate toolset: An overview of the Multivariate toolset—Help | ArcGIS Desktop 

Normally you will need training pixels and signature files to do the classification: Producing signature files, class, and cluster analysis—Help | ArcGIS Desktop 

You could use unsupervised classification, in that case your best chance would be Iso Cluster Unsupervised Classification—Help | ArcGIS Desktop . However, it may not give you the best results. There are examples of processing multi-spectral imagery use scikit-image: Image processing in Python — scikit-image  (see example here: http://remote-sensing.eu/image-classification-with-python/ )

Are you planning to process the 50 cities one by one or will you manage them in a single mosaic dataset?

[FarhanMustafa]

I am processing a single city at a time.