APPARATUS AND METHOD FOR MULTIDIMENSIONAL CHARACTERIZATION OF OBJECTS IN REAL TIME

摘要

There is disclosed herein a system for real time classification of events occurring in rapid succession and for taking some action regarding each event. In the general case data is gathered about each event (14, 16), buffered in a FIFO buffer (18), compared against a classification tree by a computer (20) and categorized. The classification decisions are stored in another FIFO buffer (34) until such time as action regarding that particular event must be taken at which time they are read (46) from buffer, interpreted and the proper signals are generated to control the action to be taken. The data that comprises the classification tree is supplied to the classification computer (20) by a user who generates it, usually using computer (38), off-line using data gathered by the data collecting apparatus of the machine. The user can use any analysis procedure he or she wishes to use and make decisions on how to classify certain classes of objects. These decisions are then codified into templates in the form of binary classification tree data. The binary classification tree data (42) is communicated to the classification computer (20) for storage and later use in classifying new data about other objects.