A hierarchical likelihood classifier with applications to remote sensing

In many classification, pattern recognition, segmentation, problems, the subspaces the union of which forms the space of interest, have some unique characteristics, and some common characteristics. The unique characteristics are expressed by one or more features. Notice that the number of features per unique characteristic is not fixed. The unique characteristics are used first to divide the space into subspaces in general larger than the target subspaces. Each subspace could contain one or more of the final or target subspaces. If it is only one the recognition is finished, else we continue with the common. Using the common features we create a maximum likelihood classifier to maximize the probability of correct classification and minimize the probability of misclassification. We apply our theory here in a remote sensing environment. The multichannel camera is mounted on an unmanned aircraft, the signals are digitized using our analog to digital hardware, compressed lossless and transmitted wireless, to a file server on earth, the camera is controled remotely.