Free flow fever screening
A method for free flow fever screening is presented. The method includes capturing a plurality of frames from thermal data streams and visual data streams related to a same scene to define thermal data frames and visual data frames, detecting and tracking a plurality of individuals moving in a free-flow setting within the visual data frames, and generating a tracking identification for each individual of the plurality of individuals present in a field-of-view of the one or more cameras across several frames of the plurality of frames. The method further includes fusing the thermal data frames and the visual data frames, measuring, by a fever-screener, a temperature of each individual of the plurality of individuals within and across the plurality of frames derived from the thermal data streams and the visual data streams, and generating a notification when a temperature of an individual exceeds a predetermined threshold temperature.
1. A method for free flow fever screening by measuring core body temperature of individuals from a distance without human intervention, the method comprising:
capturing, by one or more thermal cameras, a plurality of frames from thermal data streams and visual data streams related to a same scene to define thermal data frames and visual data frames, wherein adaptive alignment is performed for thermal data frames and the visual data frames at different depths and different positions in a same image plane;
detecting and tracking, by a person tracker, a plurality of individuals moving in a free-flow setting within the visual data frames, wherein feature points are matched across the thermal data frames and the visual data frames to determine a homography matrix used for aligning thermal/visual object pairs;
generating a tracking identification (id) for each individual of the plurality of individuals present in a field-of-view of the one or more cameras across several frames of the plurality of frames;
fusing, by a frame fuser, the thermal data frames and the visual data frames temporally and spatially, the fusing comprising detecting each individual of the plurality of individuals in the visual data frames and determining the temperature of the each individual from the thermal data frames by estimating an alignment function ƒ align by mapping from the visual data frames to the thermal data frames by (x v , y v )=ƒ align (x t , y t ), where (x v , y v ) represents a given point in the visual data frames and (x t , y t ) represents a corresponding point in the thermal data frames;
measuring, by a fever screener, a temperature of the each individual of the plurality of individuals within and across the plurality of frames derived from the thermal data streams and the visual data streams;
generating a notification when a temperature of a particular individual of the plurality of individuals exceeds a predetermined threshold temperature; and
automatically correcting for variation in measured temperatures of the plurality of individuals by dynamically adjusting parameters of the thermal camera based on detected environmental changes, and performing error correction by comparing the measured temperatures of the plurality of individuals to a reference temperature source responsive to detecting deviations from expected temperature readings.
2. The method of claim 1 , wherein a neural-network based distance compensation model is employed to correct for variation in measured temperature with distance and auto-calibration is employed to automatically correct for a drift error by using a dynamic proportional controller.
3. The method of claim 1 , wherein the temporally fusing relates to fusing in a time domain and the spatially fusing relates to fusing in a space domain.
4. The method of claim 3 , wherein the spatially fusing relates to mapping pixels between the thermal data frames and the visual data frames.
5. The method of claim 1 , wherein the tracking id is used to identify a same individual across the several frames and to record a temperature reading of the identified individual in each of the several frames to create a cluster of temperature readings for the identified individual.
6. The method of claim 1 , wherein the fever screener maintains a cache of each of the plurality of individuals with a corresponding tracking id and bounding boxes of detection, the bounding boxes defined for a body, a head, and a face of all of the plurality of individuals in each frame of the plurality of frames.
7. The method of claim 1 , wherein the fever screener measures the temperature of each individual of the plurality of individuals in a capture zone and within a region of interest.
8. The method of claim 1 , wherein the temperature measurements of each individual of the plurality of individuals within and across the plurality of frames is prioritized based on visibility of different regions of the individuals.
9. A non-transitory computer-readable storage medium comprising a computer-readable program for free flow fever screening by measuring core body temperature of individuals from a distance without human intervention, wherein the computer-readable program when executed on a computer causes the computer to perform steps of:
capturing, by one or more thermal cameras, a plurality of frames from thermal data streams and visual data streams related to a same scene to define thermal data frames and visual data frames, wherein adaptive alignment is performed for thermal data frames and the visual data frames at different depths and different positions in a same image plane;
detecting and tracking, by a person tracker, a plurality of individuals moving in a free-flow setting within the visual data frames, wherein feature points are matched across the thermal data frames and the visual data frames to determine a homography matrix used for aligning thermal/visual object pairs;
generating a tracking identification (id) for each individual of the plurality of individuals present in a field-of-view of the one or more cameras across several frames of the plurality of frames;
fusing, by a frame fuser, the thermal data frames and the visual data frames temporally and spatially, the fusing comprising detecting each individual of the plurality of individuals in the visual data frames and determining the temperature of the each individual from the thermal data frames by estimating an alignment function ƒ align by mapping from the visual data frames to the thermal data frames by (x v , y v )=ƒ align (x t , y t ), where (x v , y v ) represents a given point in the visual data frames and (x t , y t ) represents a corresponding point in the thermal data frames;
measuring, by a fever screener, a temperature of each individual of the plurality of individuals within and across the plurality of frames derived from the thermal data streams and the visual data streams;
generating a notification when a temperature of an individual of the plurality of individuals exceeds a predetermined threshold temperature; and
automatically correcting for variation in measured temperatures of the plurality of individuals by dynamically adjusting parameters of the thermal camera based on detected environmental changes, and performing error correction by comparing the measured temperatures of the plurality of individuals to a reference temperature source responsive to detecting deviations from expected temperature readings.
10. The non-transitory computer-readable storage medium of claim 9 , wherein the temporally fusing relates to fusing in a time domain and the spatially fusing relates to fusing in a space domain.
11. The non-transitory computer-readable storage medium of claim 10 , wherein the spatially fusing relates to mapping pixels between the thermal data frames and the visual data frames.
12. The non-transitory computer-readable storage medium of claim 9 , wherein the tracking id is used to identify a same individual across the several frames and to record a temperature reading of the identified individual in each of the several frames to create a cluster of temperature readings for the identified individual.
13. The non-transitory computer-readable storage medium of claim 9 , wherein the fever screener maintains a cache of each of the plurality of individuals with a corresponding tracking id and bounding boxes of detection, the bounding boxes defined for a body, a head, and a face of all of the plurality of individuals in each frame of the plurality of frames.
14. The non-transitory computer-readable storage medium of claim 9 , wherein the fever screener measures the temperature of each individual of the plurality of individuals in a capture zone and within a region of interest; and
wherein the temperature measurements of each individual of the plurality of individuals within and across the plurality of frames is prioritized based on visibility of different regions of the individuals.
15. The non-transitory computer-readable storage medium of claim 9 , wherein a neural-network based distance compensation model is employed to correct for variation in measured temperature with distance and auto-calibration is employed to automatically correct for a drift error by using a dynamic proportional controller.
16. A system for free flow fever screening by measuring core body temperature of individuals from a distance without human intervention, the system comprising:
one or more thermal cameras to capture a plurality of frames from thermal data streams and visual data streams related to a same scene to define thermal data frames and visual data frames, wherein adaptive alignment is performed for thermal data frames and the visual data frames at different depths and different positions in a same image plane;
a person tracker to detect and track a plurality of individuals moving in a free-flow setting within the visual data frames, wherein feature points are matched across the thermal data frames and the visual data frames to determine a homography matrix used for aligning thermal/visual object pairs;
a tracking identification (id) generated for each individual of the plurality of individuals present in a field-of-view of the one or more cameras across several frames of the plurality of frames;
a frame fuser configured for fusing the thermal data frames and the visual data frames temporally and spatially, the fusing comprising detecting each individual of the plurality of individuals in the visual data frames and determining the temperature of the each individual from the thermal data frames by estimating an alignment function ƒ align by mapping from the visual data frames to the thermal data frames by (x v , y v )=ƒ align (x t , y t ), where (x v , y v ) represents a given point in the visual data frames and (x t , y t ) represents a corresponding point in the thermal data frames; and
a fever-screener to measure a temperature of each individual of the plurality of individuals within and across the plurality of frames derived from the thermal data streams and the visual data streams,
wherein a notification is generated when a temperature of an individual of the plurality of individuals exceeds a predetermined threshold temperature, and variations in measured temperatures of the plurality of individuals are automatically corrected by dynamically adjusting parameters of the thermal camera based on detected environmental changes, and performing error correction by comparing the measured temperatures of the plurality of individuals to a reference temperature source responsive to detecting deviations from expected temperature readings.