Phase contrast AFM is increasingly utilized in the analysis of polymers and polymer blends. The phase contrast images potentially contain chemical information, although image interpretation can be challenging. In this study we evaluate methods for the correlation of XPS and AFM data, in order to facilitate chemical interpretation of phase contrast AFM images. Polymer grids are used to evaluate the image pre-processing required before the application of classification methods. Processing for image correlation includes resizing, image alignment and resolution matching. Considerations specific to each technique will also be discussed. Following the image processing, classification methods are used to correlate components present in the XPS and AFM images. After validating the approach on test samples, heterogeneous polymer blends are analyzed using image classification methods. This work has been partially supported by NSF ALCOM (DMR89-20147).