Abstract: The principal barrier to routine use of the ERTS multispectral scanner computer compatible tapes, rather than photointerpretation examination of the images, has been the high computing costs involved due to the large quantity of information (4 Mbytes) contained in a scene. STANSORT, the interactive program package developed at Stanford Remote Sensing Laboratories alleviates this problem, providing an extremely rapid, flexible and low cost tool for data reduction, scene classification, species searches and edge detection. The primary classification procedure, utilizing a search with variable gate widths, for similarities in the normalized, digitized spectra is described along with associated procedures for data refinement and extraction of information. The more rigorous statistical classification procedures are also explained.

Abstract: An iterative method of adaptive pattern recognition is used to allocate unclassified individuals to an a priori classification. The model is similar in form to a linear discriminant function, but the coefficient vector is determined by iteration. The method can be used with binary data, and with variables whose statistical distributions are not normal; it is therefore a useful technique for geologists.

Abstract: : The study pertains to the development of automated techniques for classification of vectorcardiograms and electrocardiograms using signal analysis techniques. A data preprocessor is developed to produce an average heartbeat from a record containing multiple heartbeats corrupted by severe baseline shifts. Several techniques for generating an ECG/VCG transformation are studied. Individual transformations are found to be quite accurate, if phase shifts among the leads are taken into account. However, patient variability appears to preclude use of a standardized transformation. The structure of the data base required to accurately test the classification algorithms is discussed and specific recommendations are made. The effect of using both two- and three-dimensional coordinate systems and a normalizing transformation for feature generation was studied; classification accuracy remained essentially unchanged.

Abstract: An algorithm to select the minimum-cost collection of binary-valued features for use with a linear pattern classifier is presented. The feature-selection algorithm is motivated by the convex-hull representation of pattern-space separability. Combinatorial analysis and linear programming are used to find the minimum-cost collection of binary-valued features associated with a given set of preclassified patterns. A description of the interaction between these algorithm components is provided. The algorithm guarantees that its optimal feature set will correctly classify every pattern in the classifier's training sample. Coinputational considerations associated with algorithm use are discussed. An application of the algorithm to a three-feature classifier is presented in detail.

Abstract: Photographic reproduction of ERTS-1 images are capable of displaying only a portion of the total information available from the multispectral scanner. Methods are being developed to generate ERTS-1 images oriented towards special users such as agriculturists, foresters, and hydrologists by applying image enhancement techniques and interactive statistical classification schemes. Spatial boundaries and linear features can be emphasized and delineated using simple filters. Linear and nonlinear transformations can be applied to the spectral data to emphasize certain ground information. An automatic classification scheme was developed to identify particular ground cover classes such as fallow, grain, rape seed or various vegetation covers. The scheme applies the maximum likelihood decision rule to the spectral information and classifies the ERTS-1 image on a pixel by pixel basis. Preliminary results indicate that the classifier has limited success in distinguishing crops, but is well adapted for identifying different types of vegetation.

Abstract: A fundamental problem in system modeling and theory is the characterization of the structure of an unknown nonlinear stochastic system when only input-output measurements are available. A method of classifying nonlinear stochastic systems, using pattern recognition and a pattern vector constructed from the input-output data, is proposed for ten stated classes of low-order nonlinear systems. The method is capable of extension to additional classes of nonlinear systems. Extensive experimental results are given to show that classification of an unknown nonlinear system, with respect to basic structural properties, can be and accomplished with a very high probability of correct classification. Various applications of the classification procedure are given, particularly in the areas of systems modeling, self-organizing control systems, and learning control systems.