Image sequence analysis has been playing an important role in many applications in computer vision, machine vision and robot vision and is also gaining interest in several fields of digital photogrammetry, especially in digital close-range photogrammetry. Examples of the application of image sequence analysis in digital close range photogrammetry are 3-D object tracking, the analysis of dynamic processes, deformation measurements, monocular or stereoscopic mapping of the environment of an autonomous robot, mobile mapping systems, biomedical motion analysis, 3-D data gathering for computer animation, and many others. Image sequence analysis techniques are also gaining interest in aerial photogrammetry in the context of automated triangulation techniques.

Sensor fusion and integration is of major importance especially in the field of mobile robots, where the information retrieved from vision systems is combined with data from other sensors like distance meters, odometers, inertial navigation systems or range images acquisition systems, but also plays a role in aerial triangulation, where INS and GPS are used as additional sources of data. The integration of rather different types of sensors requires a careful handling of redundant information and thorough error modeling. New developments of hardware components are mainly related to the storage of digital image sequences and to the development of highspeed cameras. Digital video systems now coming onto the consumer market at very reasonnable prices will considerably simplify the acquisition of monocular image sequences in the near future. The general increase of computer speed will allow realtime solutions for an increasing number of image analysis tasks in the future.

Our terms of reference are:

• Tracking single and multiple objects
• Ego-motion determination for navigation
• Object and event reconstruction from single and multiple video streams
• Camera and camera network calibration from video sequences including cameras with non-standard geometry
• Object identification from video sequences