Controlling multiple imaging sensors

A system comprising a plurality of imaging sensor nodes producing 3D structure of a scene, each sensor node having a local coordinate system and a given field of view, field of views of adjacent sensor nodes being at least partially overlapping, the nodes configured to detect locations of moving objects within the field of view of the node, and a controller, the controller configured to

• receive location data from the sensor nodes, the location data indicating the locations of the moving objects at a given time instant;
• compare the location data received from different sensor nodes at the same time instants with each other;
• performing a first determination on the basis of comparisons to obtain initial determination of which detections of different sensor nodes relate to same moving objects;
• performing a second determination on the basis of the first determination to eliminate false determinations from the initial estimation to construct a network describing the sensor network topology by determining sensor node pairs having adjacent fields of view;
• mapping the location data received from different sensor nodes to a common coordinate system;
• optimising the relationships of the fields of view of the sensor nodes with each other and the location data mapped to the common coordinate system by minimising the squared differences between the location data related to a moving object and mapped to local coordinate system and to the common coordinate system.

A method for controlling a plurality of imaging sensor nodes producing 3D structure of a scene, each sensor node having a local coordinate system and a given field of view, field of views of adjacent sensor nodes being at least partially overlapping, the method comprising:

• receiving location data from the sensor nodes, the location data indicating in the local coordinate system of the sensor node the locations of the moving objects in the field of view of the sensor node at a given time instant;
• comparing the location data received from different sensor nodes at the same time instants with each other;
• performing a first determination on the basis of comparisons to obtain initial determination of which detections of different sensor nodes relate to same moving objects;
• performing a second determination on the basis of the first determination to eliminate false determinations from the initial estimation to construct a network describing the sensor network topology by determining sensor node pairs having adjacent fields of view;
• mapping the location data received from different sensor nodes to a common coordinate system;
• and
• optimising the relationships of the fields of view of the sensor nodes with each other and the location data mapped to the common coordinate system by minimising the squared differences between the location data related to a moving object and mapped to local coordinate system and to the common coordinate system.

A non-transitory computer program product embodied on a distribution medium readable by a computer and comprising program instructions which, when loaded into an apparatus, execute a computer process comprising:

• receiving location data from a plurality of imaging sensor nodes producing 3D structure of a scene, each sensor node having a local coordinate system and a given field of view, field of views of adjacent sensor nodes being at least partially overlapping, the location data indicating in the local coordinate system of the sensor node the locations of the moving objects in the field of view of the sensor node at a given time instant;
• comparing the location data received from different sensor nodes at the same time instants with each other;
• performing a first determination on the basis of comparisons to obtain initial determination of which detections of different sensor nodes relate to same moving objects;
• performing a second determination on the basis of the first determination to eliminate false determinations from the initial estimation to construct a network describing the sensor network topology by determining sensor node pairs having adjacent fields of view;
• mapping the location data received from different sensor nodes to a common coordinate system;
• and
• optimising the relationships of the fields of view of the sensor nodes with each other and the location data mapped to the common coordinate system by minimising the squared differences between the location data related to a moving object and mapped to local coordinate system and to the common coordinate system.