The present invention relates to a hierarchical artificial neural network (HANN) for automating the recognition and identification of patterns in data matrices. It has particular, although not exclusive, application to the identification of severe storm events (SSEs) from spatial precipitation patterns, derived from conventional volumetric radar imagery. To identify characteristic features a data matrix, the data matrix is processed with a self organizing network to produce a self organizing feature space mapping. The self organizing feature space mapping is processed to produce a density characterization of the feature space mapping. The self organizing network is preferably completely unsupervised. It may, under some circumstances include a supervised layer, but it must include at least an unsupervised component for the purposes of the invention. The "self organizing feature space" is intended to include any map with the self organizing characteristics of the Kohonen Self Organizing Feature Map. The frequency vector of a CAPPI image that has been derived is a data abstraction that can be displayed directly for examination. In preferred embodiments, it is presented to a classification network, e.g. the standard CPN network, for classifying the density vector representation of the three dimensional data and displaying a representation of classified features in the three dimensional data. A novel methodology is preferably used for incorporating vigilance and conscience mechanisms in the forward counterpropagation network during training.
A method of providing weather radar images to a user includes obtaining radar image data corresponding to a weather radar image to be displayed. The radar image data is image processed to identify a feature of the weather radar image which is potentially indicative of a hazardous weather condition. The weather radar image is displayed to the user along with a notification of the existence of the feature which is potentially indicative of the hazardous weather condition. Notification can take the form of textual information regarding the feature, including feature type and proximity information. Notification can also take the form of visually highlighting the feature, for example by forming a visual border around the feature. Other forms of notification can also be used.
The computation system of the present invention comprises an improved method of moment estimation for devices which measure spectra as a function of range or time. The preferred embodiment of this system is illustrated as part of an automated meteorological monitoring system for the accurate real time detection of meteorological phenomena, such as winds, wind shear and turbulence. This automated meteorological monitoring system uses a standard weather radar transmitter to scan a predetermined volume of space with a stream of radar pulses to determine the characteristics of meteorological phenomena that are extant in the predetermined volume. The computation system of the present invention utilizes novel signal processing algorithms in the improved method of moment estimation to excise the valid data from the returns echoes, which are corrupted by the presence of contaminating signals. Separating the valid data from the noise in this manner improves the responsiveness and accuracy of the system in which this method is implemented.
A method and apparatus are provided for automatically characterizing the spatial arrangement among the data points of a time series distribution in a data processing system wherein the classification of said time series distribution is required. The method and apparatus utilize a grid in Cartesian coordinates to determine (1) the number of cells in the grid containing at least-one input data point of the time series distribution; (2) the expected number of cells which would contain at least one data point in a random distribution in said grid; and (3) an upper and lower probability of false alarm above and below said expected value utilizing a discrete binomial probability relationship in order to analyze the randomness characteristic of the input time series distribution. A labeling device also is provided to label the time series distribution as either random or nonrandom, and/or random or nonrandom.
A method and apparatus for forecasting the likely occurrence of convective weather events, such as thunderstorms. An image filter is used to identify areas of interest within a meteorological image that are likely to contain convective weather. The image filter and an image difference processor identify sub-image regions within the meteorological image that are likely to experience a growth and/or decay of weather events. The classification filter classifies sub-image regions within the meteorological image into a number of predetermined storm categories. The meteorological images are filtered using matched filters, features within the filtered images are tracked, and the resulting track vectors are combined according to the storm classification. The meteorological image, interest image, growth/decay image, classification image, and combined vectors are processed to produce the short-term forecast.
A method of clustering data includes choosing an initial cluster point and proximate data points from a set of data points. A revised data set is defined that excludes the initial cluster point and the proximate data points. A cluster point relatively far from the selected cluster point is selected. A revised data set excluding the selected cluster points and proximate data points is then defined. A decision is made whether to choose another cluster point. If so, a cluster point relatively far from the selected cluster points is selected and a revised data set excluding the selected cluster points and corresponding proximate data points is defined. If another cluster point is not to be selected, a final cluster point is chosen within the revised data set. The final cluster point is relatively far from the previously selected cluster points. The selected cluster points are then used to initiate a cluster analysis.
