Real-time feature matching in image sequences for non-structured environments. Applications to vehicle guidance

This paper describes an integrated vehicle control system with visual feedb ack. A general-purpose, low-level feature matching method, able to work in real time without any strict assumptions on the environment structure or ca mera parameters, generates low-level matching results, which are used as so urce of data for applications like mobile object tracking, among others. A generalized predictive path-tracking control approach keeps the vehicle on the trajectory defined by the moving target. In the low-level matching proc ess, block-based features (windows) are selected and tracked along a stream of monocular images; least residual square error and similarity between cl usters of features are used as constraints to select the right matching pai r between multiple candidates. Real-time performance is achieved through op timized algorithms and a parallel DSP-based multiprocessor system implement ation. Object detection and tracking is motion-based, and does not require a predefined model of the target. The integrated control system has been te sted on the ROMEO-3R experimental vehicle.