Detection, Stabilization, and Identification of Moving Objects by a Moving Observer.

摘要

The stated goal of the research was to demonstrate that robustly computable motion features can be used directly as a means of detecting and recognizing moving objects. Specifically, the goal was to design, implement, and test a general framework for detecting movement from a moving platform, and recognizing both distributed motion activity on the basis of temporal texture, and complexly moving, compact objects on the basis of their action. This recognition approach contrasts with the reconstructive approach that has typified most prior work on motion. The underlying motivation is the observation that, for objects that typically move, it is frequently easier to identify them when they are moving than when they are stationary. Specifically, in the case of temporal texture, the researchers proposed to extract statistical spatial and temporal features from approximations to the motion field and use techniques analogous to those developed for grayscale texture analysis to classify regional activities such as windblown trees, ripples on water, or chaotic fluid flow, that are characterized by complex, non-rigid motion. For action identification, they proposed to use the spatial and temporal arrangement of motion features in conjunction with simple geometric image analysis to identify complexly moving objects such as machinery and locomoting people and animals. The proposed work has practical applications in monitoring and surveillance, and as a component of a sophisticated visual system.