Method and system for video-based vehicle tracking adaptable to traffic conditions
View Patent ↗A method and system for adaptable video-based object tracking includes acquiring video data from a scene of interest and identifying an initial instance of an object of interest in the acquired video data. A representation of a target object is then established. One or more motion parameters associated with said scene of interest are used to adjust the size of a search neighborhood associated with said target object. The target object is then tracked frame-by-frame in the video data.
1. A method for video-based object tracking, said method comprising:
selecting at least one motion parameter from a plurality of motion parameters associated with a scene of interest;
defining a motion environment state according to said selected at least one motion parameter, wherein said selected at least one motion parameter is indicative of a pattern of motion in said scene of interest;
setting a size of a search neighborhood associated with a target object according to said motion environment state;
adjusting said motion environment state as said pattern of motion in said scene of interest changes;
adjusting said size of said search neighborhood associated with said target object in response to said adjusted motion environment state; and
determining a frame-by-frame location of said target object using said adjusted search neighborhood size in order to track said target object in said video data.
2. The method of claim 1 wherein said steps of selecting said at least one motion parameter and adjusting said size of said search neighborhood are applied separately to each of a plurality of motion directions associated with said scene of interest.
3. The method of claim 1 wherein said at least one of said plurality of motion parameters associated with said scene of interest is selected according to at least one of a pattern of motion associated with said scene and a plurality of scene characteristics associated with said scene.
4. The method of claim 3 wherein determining a frame-by-frame location of said target object further comprises applying at least one of:
template matching tracking;
point tracking;
mean shift tracking; and
particle filter tracking.
5. The method of claim 4 wherein selecting at least one motion parameter from said plurality of motion parameters associated with said scene of interest, further comprises continuously reviewing and updating said selected at least one motion parameter according to a relative influence of said at least one motion parameter associated with said scene.
6. The method of claim 4 wherein said plurality of motion parameters comprises at least one of:
a traffic volume;
a time of day;
a day of the week;
a specific date;
a weather condition;
an incident report;
a vehicle count;
a direction of motion;
a location of motion;
a speed; and
an acceleration.
7. The method of claim 1 further comprising extracting an initial instance of an object of interest in acquired video data by applying at least one of:
a double difference technique and a morphological filtering;
a background subtraction technique and a morphological filtering; and
a computer vision technique.
8. The method of claim 1 wherein said search neighborhood is configured to be spatially dependent in order to account for at least one characteristic of said target object resulting from at least one of a perspective and a distortion of a camera collecting video data of said scene of interest.
9. The method of claim 1 wherein establishing a representation of a target object from said initial instance of said object of interest comprises applying at least one of:
a color features representation technique;
a texture appearance technique;
a gradient magnitude technique;
a Harris Corners technique;
a feature transform technique;
a SIFT technique; and
a SURF technique.
10. A method for video-based object tracking, said method comprising:
selecting at least one motion parameter from a plurality of motion parameters associated with a scene of interest, wherein said selected at least one motion parameter is selected according to at least one of a pattern of motion associated with said scene of interest and a plurality of scene characteristics associated with said scene of interest;
defining a motion environment state according to said selected at least one motion parameter;
continuously reviewing and updating said motion environment state according to a relative influence of said selected at least one motion parameter on said pattern of motion associated with said scene of interest;
adjusting a size of a search neighborhood associated with a target object in response to said motion environment state; and
determining, based on said adjusted search neighborhood size, a frame-by-frame location of said target object in order to track said target object in said video data.
11. The method of claim 10 wherein said steps of selecting said at least one motion parameter, continuously reviewing and updating said selected at least one motion parameter, and adjusting said size of said search neighborhood are applied separately to each of a plurality of motion directions associated with said scene of interest.
12. The method of claim 11 wherein determining a frame-by-frame location of said target object further comprise at least one of:
template matching tracking;
point tracking;
mean shift tracking; and
particle filter tracking.
13. The method of claim 11 wherein said plurality of motion parameters comprises at least one of:
a traffic volume;
a time of day;
a day of the week;
a specific date;
a weather condition;
an incident report;
a vehicle count;
a direction of motion;
a location of motion;
a speed; and
an acceleration.
14. The method of claim 10 further comprising extracting an initial instance of an object of interest in acquired video data by applying at least one of:
a double difference technique and a morphological filtering;
a background subtraction technique and a morphological filtering; and
a computer vision technique.
15. A video-based object tracking system, comprising:
a video camera configured to collect video data;
a processor; and
a computer-usable medium embodying computer code, said computer-usable medium being coupled to said processor, said computer code comprising non-transitory instruction media executable by said processor configured for:
selecting at least one motion parameter from a plurality of motion parameters associated with a scene of interest;
defining a motion environment state according to said selected at least one motion parameter, wherein said selected at least one motion parameter is indicative of a pattern of motion in said scene of interest;
setting a size of a search neighborhood associated with said target object according to said motion environment state;
continuously reviewing and updating said motion environment state according to a relative influence of said selected at least one motion parameter associated with said scene of interest;
adjusting said size of said search neighborhood associated with said target object in response to said updated motion environment state; and
determining, based on said adjusted search neighborhood size, a frame-by-frame location of said target object in order to track said target object in said video data.
16. The system of claim 15 wherein said at least one of said plurality of motion parameters associated with said scene of interest is selected according to at least one of a pattern of motion associated with said scene and a plurality of scene characteristics associated with said scene.
17. The system of claim 16 wherein said instructions for determining a frame-by-frame location of said target object further comprise instructions configured for at least one of:
template matching tracking;
point tracking;
mean shift tracking; and
particle filter tracking.
18. The system of claim 16 wherein said plurality of motion parameters comprises at least one of:
a traffic volume;
a time of day;
a day of the week;
a specific date;
a weather condition;
an incident report;
a vehicle count;
a direction of motion;
a location of motion;
a speed; and
an acceleration.
19. The system of claim 15 further comprising instructions configured for extracting an initial instance of an object of interest in acquired video data by applying at least one of:
a double difference technique and a morphological filtering;
a background subtraction technique and a morphological filtering; and
a computer vision technique.
20. The system of claim 15 wherein said instructions for establishing a representation of a target object from said initial instance of said object of interest comprises instructions configured for applying at least one of:
a color features representation technique;
a texture appearance technique;
a gradient magnitude technique;
a Harris Corners technique;
a feature transform technique;
a SIFT technique; and
a SURF technique.