Abstract: The invention provides methods, systems and computer program products for image based detection of occurrence of an anomalous event within a process environment. Detection of occurrence of an anomalous event comprises (i) receiving a first set of information from a first image acquisition device, (ii) analyzing the first set of information for determining whether the first image frame images an occurrence of an anomalous event, (iii) receiving a second set of information generated at a second device, wherein the second set of information represents a state of the process environment, (iv) analyzing the second set of information for determining whether an anomalous event has occurred, and (v) generating an anomaly identification decision based at least on output from analysis of the second set of information.

Abstract: A method for anomaly detection in process control environments is provided. The method includes receiving a first information from first field devices located within a process control environment, and transmitting a first vector based on the first information, to a first detector implemented within an edge node that provides gateway access to the environment. The first detector implements a first classifier model. When the first detector detects an anomaly event based on the first vector, a second vector is transmitted to a second detector implemented within a cloud based server. The second vector, different from the first vector, is generated based on second information from second field devices located within the process control environment. The second detector implements a second classifier model different from the first classifier model. An anomaly event alert is generated when one of the first detector and the second detector detects an anomaly event.

Abstract: The invention provides methods, systems and computer program products for image based detection of occurrence of an anomalous event within a process environment. Detection of occurrence of an anomalous event comprises (i) receiving a first set of information from a first image acquisition device, (ii) analyzing the first set of information for determining whether the first image frame images an occurrence of an anomalous event, (iii) receiving a second set of information generated at a second device, wherein the second set of information represents a state of the process environment, (iv) analyzing the second set of information for determining whether an anomalous event has occurred, and (v) generating an anomaly identification decision based at least on output from analysis of the second set of information.

Abstract: The invention provides systems and methods for detection of anomaly events or error events within a process environment by implementing a hybrid of centralized classifier models and edge node based classifier models.