초록 ▼

Enhanced portal security is provided through stereoscopy, including a stereo door sensor for detecting and optionally preventing access violations, such as piggybacking and tailgating. A portal security system can include a 3D imaging system that generates a target volume from plural 2D images of a

Enhanced portal security is provided through stereoscopy, including a stereo door sensor for detecting and optionally preventing access violations, such as piggybacking and tailgating. A portal security system can include a 3D imaging system that generates a target volume from plural 2D images of a field of view about a portal; and a processor that detects and tracks people candidates moving through the target volume to detect a portal access event.

대표청구항 ▼

What is claimed is: 1. A portal security system, comprising: a sensor capturing plural 2D images of a field of view about a portal; a 3D imaging system generating a target volume from the plural 2D images of the field of view about the portal, the portal defining a moving volume through which a per

What is claimed is: 1. A portal security system, comprising: a sensor capturing plural 2D images of a field of view about a portal; a 3D imaging system generating a target volume from the plural 2D images of the field of view about the portal, the portal defining a moving volume through which a person passes from an entrance to an exit; and a processor dynamically varying a position of the target volume within the field of view as the moving volume moves through the field of view, the processor varying the position to correspond to the moving volume according to a position of the portal, the processor detecting and tracking people candidates moving through the target volume to detect a portal access event. 2. The portal security system of claim 1 wherein the target volume is less than the field of view. 3. The portal security system of claim 1, wherein the target volume has a depth less than a depth associated with the field of view. 4. The portal security system of claim 1 wherein the portal access event is a piggyback event. 5. The portal security system of claim 1 wherein the portal access event is a person event that triggers a tailgating event. 6. The portal security system of claim 1 wherein the portal access event is a person event that activates the portal. 7. The portal security system of claim 1 wherein the portal access event is an ambiguous event. 8. The portal security system of claim 1, further comprising: plural cameras receiving plural 2D images of a field of view about a portal a 3D image generator generating a 3D model from the plural 2D images; a filter providing a target volume from the 3D model; a people candidate detector detecting people candidates within the target volume; and a people candidate tracker tracking movement of the people candidates over time to detect portal access events. 9. The portal security system of claim 1 where the portal is a revolving door. 10. The portal security system of claim 1, wherein the processor automatically calibrates a 3D coordinate system that corresponds to a 3D coordinate system of a portal scene. 11. A portal security system, comprising: a 3D imaging system generating an entry target volume and an exit target volume, the entry target volume being generated from plural 2D images of a field of view about an entry of a portal, the exit target volume being generated from plural 2D images of a field of view about an exit of the portal, at least a subset of the exit target volume being locally distinct from the entry target volume; and a processor dynamically varying position of the entry target volume and the exit target volume within the respective fields of view to correspond to respective moving volumes defined by the portal, the respective moving volumes moving through the respective fields of view, the processor detecting and tracking people candidates moving through the entry target volume and the exit target volume to detect a portal access event. 12. A method of providing portal security, comprising: capturing plural 2D images of a field of view about a portal; generating a target volume from the plural 2D images of the field of view about the portal, the portal defining a moving volume through which a person passes from an entrance to an exit; dynamically varying a position of a target volume within the field of view as the moving volume moves through the field of view, a processor varying the position to correspond to the moving volume according to a position of the portal, detecting people candidates within the target volume; and tracking the people candidates moving through the target volume to detect a portal access event. 13. The method of claim 12 wherein the target volume is less than the field of view. 14. The method of claim 12, wherein the target volume has a depth less than a depth associated with the field of view. 15. The method of claim 12 wherein the portal access event is a piggyback event. 16. The method of claim 12 wherein the portal access event is a person event that triggers a tailgating event. 17. The method of claim 12 wherein the portal access event is a person event that activates the portal. 18. The method of claim 12 wherein the portal access event is an ambiguous event. 19. The method of claim 12, further comprising: receiving plural 2D images of a field of view about a portal; generating a 3D model from the 2D images; providing a target volume from the 3D model; detecting people candidates within the target volume; and tracking movement of the people candidates over time to detect portal access events. 20. The method of claim 12 wherein the portal is a revolving door, a sliding door, a swinging door, or a man trap. 21. The method of claim 12 wherein detecting people candidates comprises generating a low resolution representation having a topological profile of the target volume; identifying a location of a peak within the low resolution representation; and converting the location of the peak into an approximate location of a people candidate in a high resolution representation of the target volume. 22. The method of claim 12 wherein detecting people candidates comprises generating a head template having dimensions corresponding to a height and an approximate location of a people candidate; and determining a fine location of the people candidate by matching the height of the people candidate to heights within an area of a representation of the target volume the area corresponding to the dimensions of the head template. 23. The method of claim 22, further comprising: determining a distribution of heights within the area corresponding to the dimensions of the head template and the fine location of the people candidate; and discarding the people candidate if the distribution of heights covers less than a height distribution for a head of a person. 24. The method of claim 22, further comprising: determining plural fine locations for plural people candidates; comparing the plural fine locations in three dimensional space; discarding at least one of the plural people candidates having a fine location that overlaps with another fine location of another people candidate. 25. The method of claim 12, wherein tracking the people candidates comprises: generating confidence scores corresponding to numbers of people candidates being within the target volume; and detecting the portal access event from a series of the generated confidence scores. 26. The method of claim 25, further comprising generating confidence scores corresponding to at least one person being within the target volume by: generating a target volume in three dimensional space; generating individual scores for three dimensional points within the target volume, each of the individual scores corresponding to a height of one of the three dimensional points relative to a height threshold; generating a cumulative score from the individual scores; and computing a confidence score corresponding to at least one person being within the target volume based on the cumulative score and a second threshold. 27. The method of claim 25 further comprising generating confidence scores corresponding to two or more people being within the target volume based on a separation between head locations of the people candidates. 28. The method of claim 25 further comprising generating confidence scores corresponding to two or more people being within the target volume based on a template score generated from matching head templates to head locations of the people candidates. 29. The method of claim 12 further comprising: automatically calibrating a 3D coordinate system that corresponds to a 3D coordinate system of a portal scene. 30. The method of claim 12, further comprising controlling access to the portal according to the portal access event. 31. The method of claim 12, further comprising reporting an alert based on the portal access event. 32. A method of providing portal security, comprising: generating an entry target volume from plural 2D images of a field of view about an entry of a portal; generating an exit target volume from plural 2D images of a field of view about an exit of the portal, at least a subset of the exit target volume being locally distinct from the entry target volume; dynamically varying position of the entry target volume and the exit target volume within the respective fields of view to correspond to respective moving volumes defined by the portal, the respective moving volumes moving through the respective fields of view; detecting people candidates within the entry target volume and the exit target volume; and tracking the people candidates moving through the entry target volume and the exit target volume to detect a portal access event. 33. The method of claim 32, further comprising controlling access to the portal according to the portal access event. 34. The method of claim 32, further comprising reporting an alert based on the portal access event. 35. A method of providing portal security, comprising: generating a target volume from plural 2D images of a field of view about a portal; and detecting people candidates within the target volume; tracking the people candidates moving trough the target volume to detect a portal access event, wherein tracking the people candidates comprises: generating confidence scores corresponding to numbers of people candidates being within the target volume; and detecting the portal access event from a series of the generated confidence scores; and generating confidence scores corresponding to only one person being within the target volume by: generating a target volume in three dimensional space; removing a portion of the target volume that corresponds to one of the people candidates in the target volume; generating individual scores for three dimensional points remaining within the target volume, each of the individual scores corresponding to a height of one of the remaining three dimensional points relative to a height threshold; generating a cumulative score from the individual scores; and computing a confidence score corresponding to only one person being within the target volume based on the cumulative score and a second threshold. 36. A method of providing portal security, comprising: generating a target volume from plural 2D images of a field of view about a portal, the portal defining a moving volume through which a person passes from an entrance to an exit; dynamically varying a position of a target volume to correspond to the moving volume according to a position of the portal; detecting people candidates within the target volume; and tracking the people candidates moving through the target volume to detect a portal access event; wherein tracking the people candidates comprises: generating confidence scores corresponding to numbers of people candidates being within the target volume; detecting the portal access event from a series of the generated confidence scores; generating a set of traces over a sequence of frames, each trace in the set corresponding to a trajectory of a people candidate during the sequence of frames; assigning weights to the sequence of frames, such that the weights are a uniform weight unless the set of traces includes a first trace and a second trace with the second trace being generated over a set of frames that is substantially less than the first trace, the weights assigned to the set of frames that generated the second trace being assigned a weight less than the uniform weight; and applying the assigned weights of the sequence of frames to the confidence scores of corresponding frames.