Many approaches for object detection are based on Markov Random Field (MRF) and Conditional Random Field (CRF) respectively. MRF and CRF have very different characteristics. This work discusses in detail their strength and weaknesses. From the discussion, a new object detection algorithm using combination of CRF and MRF was derived. We utilize the algorithm to detect urban areas, and corresponding to the urban area object, we introduce a generic feature vector for each image site. The proposed algorithm was tested extensively on a large number of remote sensing images, and very promising results can be presented.