Classifying land development in high resolution satellite images using straight line statistics

We introduce a set of measures based on straight lines to assess land development levels in high resolution (1 meter) satellite images. Urban areas exhibit a preponderance of straight line features, generally appearing in fairly simple, quasiperiodic organizations. Wilderness and rural areas produce line structures in more random spatial arrangements. We use this observation to perform an initial triage on the image to restrict the attention of subsequent, more computationally intensive analyses. We extract statistical measures based on straight lines to guide the analysis. We base these measures on orientation, length, contrast, periodicity and location. We trained and tested parametric and non-parametric classifiers using the feature set. Finally, we introduce a decision system performing region classification via an overlapped voting method for consensus discovery.