System and method for identifying an object in an image

1. A method for processing image data, the method comprising:

identifying a background scene in a background frame of an image;

setting each of a plurality of pixel locations in the background frame as a respective background pixel;

for each of the pixel locations, determining a respective running mean and a respective running variance based on a brightness component of the background frame at the pixel location;

identifying a subsequent frame as a current frame;

determining a mean weight associated with the current frame based on exposure settings of an image capturing device significantly changing;

for each of the pixel locations in the current frame identified as a background pixel:

determining a pixel mean weight based on whether one of the current pixel location changed in the current frame from a foreground pixel to a background pixel and the current pixel location did not change in the current frame from a foreground pixel to a background pixel;

updating the running mean based on one of the mean weight and the pixel mean weight;

updating the running variance based on one of a plurality of variance weights;

for each of the pixel locations in the current frame:

identifying if the current pixel location is one of a background pixel and a foreground pixel based on a brightness component of the current pixel location in the current frame, a current running mean of the current pixel location, and a current running variance of the current pixel location; and

updating the running mean for the current pixel locations identified as a foreground pixel.

2. The method for processing image data as set forth in claim 1 , further including:

repeating the step of identifying a subsequent frame as a current frame along with the steps of updating the running mean, and updating the running variance for each of the pixel locations in the current frame.

3. A method for processing image data, the method comprising:

identifying a background scene in a background frame of an image;

setting each of a plurality of pixel locations in the background frame as a respective background pixel;

for each of the pixel locations, determining a respective running mean and a respective running variance based on a brightness component of the background frame at the pixel location;

identifying a subsequent frame as a current frame;

determining a variance weight associated with the current frame based on exposure settings of an image capturing device significantly changing;

for each of the pixel locations in the current frame identified as a background pixel:

determining a pixel variance weight based on an exposure associated with the current frame;

updating the running mean based on one of a plurality of mean weights;

updating the running variance based on one of the variance weight and the pixel variance weight;

for each of the pixel locations in the current frame:

identifying if the current pixel location is one of a background pixel and a foreground pixel based on a brightness component of the current pixel location in the current frame, a current running mean of the current pixel location, and a current running variance of the current pixel location.

4. The method for processing image data as set forth in claim 3 , wherein the step of updating the running mean includes updating the running mean according to:

RM Pixel=((1−Weight M )*Prev RM Pixel)+(Weight M *Pixel BC ),

where: RMPixel is the updated running mean for the current pixel location in the current frame;

WeightM is a current mean weight for the current pixel location in the current frame;

PrevRMPixel is the running mean of the current pixel location from a previous frame; and

PixelBC is the brightness component of the current pixel location in the current frame.

5. The method for processing image data as set forth in claim 3 , wherein updating the running variance based on the variance weight includes updating an instantaneous variance according to:

PixelVar=(Pixel BC−RM Pixel) 2

where: PixelBC is the brightness component of the current pixel location; and

RMPixel is the current running mean value at the current pixel location.

6. The method for processing image data as set forth in claim 5 , wherein the step of updating the running variance based on a current pixel variance weight includes updating the running variance according to:

RV Pixel=((1−Weight V )*Prev RV Pixel)+(Weight V *PixelVar),

where: RVPixel is the updated running variance for the current pixel location in the current frame;

WeightV is the running variance for the current pixel location in the current frame;

PrevRVPixel is the running variance of the current pixel location from a previous frame; and

PixelVar is an instantaneous variance of the brightness component of the current pixel location with respect to a background value.

7. The method for processing image data as set forth in claim 3 , wherein the step of identifying if the current pixel location is one of a background pixel and a foreground pixel includes identifying if the current pixel location is one of a background pixel and a foreground pixel according to:

diff 2 =(Pixel BC−RM Pixel) 2 ,

where: PixelBC is the brightness component of the current pixel location in the current frame; and

RMPixel is the updated running mean for the current pixel location in the current frame; and

identifying if the current pixel location is one of a background pixel and a foreground pixel based on diff 2 .

8. The method for processing image data as set forth in claim 7 , wherein the step of identifying if the current pixel location is one of a background pixel and a foreground pixel includes identifying if the current pixel location is the foreground pixel if:

diff 2 >a predetermined threshold; and

diff 2 >(k*RVPixel),

where: RVPixel is the updated running variance for the current pixel location in the current frame; and

k is a constant.

9. The method for processing image data as set forth in claim 3 , further including:

identifying a bounding box surrounding a group of the foreground pixel locations;

identifying each of the pixel locations within the bounding box as a foreground pixel for updating the running variance.

10. The method for processing image data as set forth in claim 9 , further including:

defining the bounding box as any geometric shape.