The neural control system of a high speed monocular camera head for the tracking of real-world targets is presented in this paper. The tracking system consists of four subsystems: monocular camera head. adaptive image processing system for estimation of the momentary position of object, neural network predictor and PID-controller, controlling motors of the camera head. The designed neural network tracking system performs smooth pursuit of slow objects (50°) with a foveal error less than 0.7° and is able to track objects up to a maximum speed of 320° with foveal error less than 4.5°.

An image processing system for the real-time object detection and recognition was designed on the principles of active vision and sequential analysis. The real-world visual tasks can be solved due to predictive control of the vision sensor. The sequential analysis allows real-time implementation of the system on low cost DSP hardware. The system was implemented on the DSP TMS320C50 and requires 18-30 ms for the detection and recognition of the object.

In this paper(1) we name and describe some necessary features that active vision systems realize to enable industrial applications. Based on capabilities that one can find in biological systems (especially primate oculomotion) and on their technical implementation by neural computation, novel tasks come into reach of beeing suitable for applications.Keeping the object of interest within the center of the sensitive optical area by changing position and gaze of the optical sensor, is the most challenging feature of active visual systems. A pan-tilt unit carrying a camera can be enhanced for beeing an adequate active vision system when controlled by a neural network. With these actual state of developments and results, novel tasks come into reach of beeing realized, suitable for industrial applications.