3D imaging using structured light for accurate vehicle occupancy detection
View Patent ↗What is disclosed is a method which combines structured illumination in the SWIR wavelength range with the detection capabilities of NIR to generate a 3D image of a scene for accurate vehicle occupancy determination. In one embodiment, structured light is projected through a customized optical element comprising a patterned grid. Wavelengths of the received structured pattern are shifted to a CCD detectable range. The shifted light comprises an image in a structured pattern. The wavelength-shifted light is detected using an infrared detector operating in the NIR. For each pixel in the detected patterned image, an amount of distortion caused by 3D surface variation at this pixel location is determined. The distortion is converted to a depth value. The process repeats for all pixels. A 3D image is constructed using each pixel's depth value. The number of occupants in the vehicle is determined from the constructed 3D image.
1. A method for 3D imaging of an occupant in a motor vehicle for vehicle occupancy determination, the method comprising:
projecting an illumination light source onto a motor vehicle intended to be analyzed for vehicle occupancy determination, said illumination source comprising structured light in the Short Wave Infrared (SWIR) band, said structured light being projected through an optical element comprising a patterned grid, a reflection of said source light being sensed by an imaging sensor;
capturing said projected pattern by said imaging sensor;
calibrating said captured projected pattern comprising:
determining a relationship between actual intensity of said projected pattern and pixel intensity values;
fitting said established relationship with a high order polynomial function;
calculating an inverse function; and
using said inverse function to rectify said pattern;
shifting sensed reflected source light of said captured rectified pattern to a wavelength range of a charge-coupled device (CCD), said wavelength-shifted light being an image in a structured pattern;
using an infrared detector operating in the near infrared range of 700-1000 nm to detect said wavelength-shifted light;
for each pixel in said structured patterned image:
analyzing said detected wavelength-shifted light to determine an amount of distortion caused by 3D surface variation at this pixel location; and
converting said distortion to a depth value for this pixel;
constructing a 3D image using each pixel's respective depth value; and
determining a number of occupants in said vehicle by analyzing said 3D image.
2. The method of claim 1 , wherein said Short Wave Infrared band is tuned to an available wavelength converter range of 1495 to 1550 nm.
3. The method of claim 1 , wherein wavelengths of said pattern are converted to a CCD detectable range using a phosphor coated converter which shifts the received light to a Si-based CCD detectable wavelength range.
4. The method of claim 1 , wherein, in order to enhance a contrast, an intensity curve of said illumination projector is altered with a gamma transformation.
5. The method of claim 1 , further comprising calibrating said light source to recover the linearity of illumination intensity.
6. The method of claim 1 , further comprising determining whether a traffic violation has occurred based upon said determined number of occupants.
7. The method of claim 6 , further comprising alerting an authority in the instance wherein a traffic violation has occurred.
8. A system for 3D imaging of an occupant in a motor vehicle for vehicle occupancy determination, the system comprising:
an illumination light source comprising structured light in the Short Wave Infrared (SWIR) band, said structured light being projected through an optical element comprising a patterned grid;
an imaging sensor for sensing a reflection of said light source;
an infrared detector operating in the near infrared range of 700-1000 nm; and
a processor in communication with a memory for executing machine readable instructions for performing:
capturing said projected pattern by said imaging sensor;
calibrating said captured projected pattern comprising:
determining a relationship between actual intensity of said projected pattern and pixel intensity values;
fitting said established relationship with a high order polynomial function;
calculating an inverse function; and
using said inverse function to rectify said pattern;
shifting sensed reflected source light of said captured rectified pattern to a wavelength range of a charge-coupled device (CCD), said wavelength-shifted light being an image in a structured pattern;
using an infrared detector operating in the near infrared range of 700-1000 nm to detect said wavelength-shifted light;
for each pixel in said structured patterned image:
analyzing said detected wavelength-shifted light to determine an amount of distortion caused by 3D surface variation at this pixel location; and
converting said distortion to a depth value for this pixel;
constructing a 3D image using each pixel's respective depth value; and
determining a number of occupants in said vehicle by analyzing said 3D image.
9. The system of claim 8 , wherein said Short Wave Infrared band is tuned to an available wavelength converter range of 1495 to 1550 nm.
10. The system of claim 8 , wherein wavelengths of said pattern are converted to a CCD detectable range using a phosphor coated converter which shifts the received light to a Si-based CCD detectable wavelength range.
11. The system of claim 8 , wherein, in order to enhance a contrast, an intensity curve of said illumination projector is altered with a gamma transformation.
12. The system of claim 8 , further comprising calibrating said light source to recover the linearity of illumination intensity.
13. The system of claim 8 , further comprising determining whether a traffic violation has occurred based upon said determined number of occupants.
14. The system of claim 13 , further comprising alerting an authority in the instance wherein a traffic violation has occurred.