Determining a pixel classification threshold for vehicle occupancy detection
View Patent ↗What is disclosed is a system and method for determining a pixel classification threshold for vehicle occupancy determination. An IR image of a moving vehicle is captured using a multi-band IR imaging system. A driver's face is detected using a face recognition algorithm. Multi-spectral information extracted from pixels identified as human tissue of the driver's face is used to determine a pixel classification threshold. This threshold is then used to facilitate a classification of pixels of a remainder of the IR image. Once pixels in the remainder of the image have been classified, a determination can be made whether the vehicle contains additional human occupants other than the driver. An authority is alerted in the instance where the vehicle is found to be traveling in a HOV/HOT lane requiring two or more human occupants and a determination has been made that the vehicle contains an insufficient number of human occupants.
1. A method for determining a threshold for pixel classification in a vehicle occupancy detection system, the method comprising:
receiving an IR image of a moving vehicle intended to be analyzed for passenger occupancy in the vehicle, said image having been captured using a multi-band IR imaging system comprising an IR detector and an IR Illuminator, said imaging system having collected intensity values for each pixel in said image;
detecting pixels of human tissue of a driver of said vehicle in said IR image; and
using multi-spectral information of said detected human tissue pixels to determine a pixel classification threshold C th comprising:
C th =C m −βV c ,
where C m is a mean of correlation coefficient values of pixels of said driver's detected face, V c is a variance of said correlation, and β is a constant.
2. The method of claim 1 , further comprising using said pixel classification threshold to classify human tissue pixels in a remainder of said image, said pixels being classified according to a correlation method comprising:
C
=
∑
i
=
1
i
=
N
[
I
cm
(
i
)
]
[
I
cs
(
i
)
]
∑
i
=
1
i
=
N
[
I
cm
(
i
)
]
2
∑
i
=
1
i
=
N
[
I
cs
(
i
)
]
2
,
where I cm (i) is a measured intensity of pixels of said passenger, I cs (i) is a facial intensity of the said driver, and N is the total number of wavelength band of said multi-band IR imaging system, said classification being based upon said correlation being larger or smaller than said threshold C th .
3. The method of claim 1 , further comprising:
clipping said image to a windshield area of said vehicle; and
analyzing said windshield area to determine an approximate location of a driver's side and a passenger side of said vehicle's front passenger compartment.
4. The method of claim 1 , further comprising determining a number of human occupants in said vehicle based upon said pixel classifications.
5. The method of claim 4 , further comprising alerting an authority in the instance where a front passenger seat is determined to not be occupied by a human occupant and said vehicle is traveling in a HOV/HOT lane.
6. A system for determining a threshold for pixel classification in a vehicle occupancy detection system, the system comprising:
a multi-band IR imaging sensor for capturing an IR image of a moving vehicle intended to be analyzed for passenger occupancy in the vehicle, said capture IR image comprising intensity values collected for each pixel in said image; and
a processor in communication with said imaging sensor and a memory, said processor executing machine readable instructions for performing:
detecting pixels of human tissue of a driver of said vehicle in said IR image; and
using multi-spectral information of said detected human tissue pixels to determine a pixel classification threshold C th comprising:
C th =C m −βV c ,
where C m , is a mean of correlation coefficient values of pixels of said driver's detected face, V c is a variance of said correlation, and β is a constant.
7. The system of claim 6 , further comprising using said pixel classification threshold to classify human tissue pixels in a remainder of said image, said pixels being classified according to a correlation method comprising:
C
=
∑
i
=
1
i
=
N
[
I
cm
(
i
)
]
[
I
cs
(
i
)
]
∑
i
=
1
i
=
N
[
I
cm
(
i
)
]
2
∑
i
=
1
i
=
N
[
I
cs
(
i
)
]
2
,
where I cm (i) is a measured intensity of pixels of said passenger, I cs (i) is a facial intensity of the said driver, and N is the total number of wavelength band of said multi-band IR imaging system, said classification being based upon said correlation being larger or smaller than said threshold C th .
8. The system of claim 6 , further comprising:
clipping said image to a windshield area of said vehicle; and
analyzing said windshield area to determine an approximate location of a driver's side and a passenger side of said vehicle's front passenger compartment.
9. The system of claim 6 , further comprising determining a number of human occupants in said vehicle based upon said pixel classifications.
10. The system of claim 9 , further comprising alerting an authority in the instance where a front passenger seat is determined to not be occupied by a human occupant and said vehicle is traveling in a HOV/HOT lane.
11. A method for determining a threshold for pixel classification in a vehicle occupancy detection system, the method comprising:
receiving an IR image of a moving vehicle intended to be analyzed for passenger occupancy in the vehicle, said image having been captured using a multi-band IR imaging system comprising an IR detector and an IR Illuminator, said imaging system having collected intensity values for each pixel in said image;
detecting pixels of human tissue of a driver of said vehicle in said IR image; and
using multi-spectral information of said detected human tissue pixels to determine a pixel classification threshold S th comprising:
S th =S m −γV s ,
where S m is a mean of intensity ratio values of pixels of said driver's detected face, V s is a variance of said values, and γ is a constant.
12. The method of claim 11 , further comprising using said pixel classification threshold to classify human tissue pixels in a remainder of said image, said pixels being classified according to a ratio method comprising:
I
cm
(
i
)
I
cm
(
j
)
,
where i,j are any N-band indices different from each other, said classification being based upon whether said ratio is larger or smaller than said threshold.
13. The method of claim 11 , further comprising:
clipping said image to a windshield area of said vehicle; and
analyzing said windshield area to determine an approximate location of a driver's side and a passenger side of said vehicle's front passenger compartment.
14. The method of claim 11 , further comprising determining a number of human occupants in said vehicle based upon said pixel classifications.
15. The method of claim 11 , further comprising alerting an authority in the instance where a front passenger seat is determined to not be occupied by a human occupant and said vehicle is traveling in a HOV/HOT lane.
16. A system for determining a threshold for pixel classification in a vehicle occupancy detection system, the system comprising:
a multi-band IR imaging sensor for capturing an IR image of a moving vehicle intended to be analyzed for passenger occupancy in the vehicle, said capture IR image comprising intensity values collected for each pixel in said image; and
a processor in communication with said imaging sensor and a memory, said processor executing machine readable instructions for performing:
detecting pixels of human tissue of a driver of said vehicle in said IR image; and
using multi-spectral information of said detected human tissue pixels to determine a pixel classification threshold S th comprising:
S th =S m −γV s ,
where S m is a mean of intensity ratio values of pixels of said driver's detected face, V s is a variance of said values, and γ is a constant.
17. The system of claim 16 , further comprising using said pixel classification threshold to classify human tissue pixels in a remainder of said image, said pixels being classified according to a ratio method comprising:
I
cm
(
i
)
I
cm
(
j
)
,
where i,j are any N-band indices different from each other, said classification being based upon whether said ratio is larger or smaller than said threshold.
18. The system of claim 16 , further comprising:
clipping said image to a windshield area of said vehicle; and
analyzing said windshield area to determine an approximate location of a driver's side and a passenger side of said vehicle's front passenger compartment.
19. The system of claim 16 , further comprising determining a number of human occupants in said vehicle based upon said pixel classifications.
20. The system of claim 16 , further comprising alerting an authority in the instance where a front passenger seat is determined to not be occupied by a human occupant and said vehicle is traveling in a HOV/HOT lane.