Occupancy detection for managed lane enforcement based on localization and classification of windshield images

• US 8,824,742 B2
• Filed: 06/19/2012
• Issued: 09/02/2014
• Est. Priority Date: 06/19/2012
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Skaff, Sandra, Xu, Beilei, Paul, Peter, Saunders, Craig, Perronnin, Florent
• Primary Examiner(s): SHAH, UTPAL D
• Application Number:
  US13/527,228
• Publication Number:
  US 20130336538A1
• Time in Patent Office: 805 Days
• Field of Search: 382/104, 382/224
• US Class Current: 382/104
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]
• CPC Class Codes: G06V 20/54 of traffic, e.g. cars on the road, trains or boats
• CPC Class Codes: G06V 20/59 inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions

Abstract

A system for detecting a vehicle occupancy violation includes an image capture module that acquires an image including a vehicle cabin from a camera positioned to view oncoming traffic. The system includes a violation determination device, which includes a feature extraction module that processes the image pixels for determining an image descriptor. The process is selected from a group consisting of a Successive Mean Quantization Transform; a Scale-Invariant Feature Transform; a Histogram of Gradients; a Bag-of-Visual-Words Representation; a Fisher Vector Representation; and, a combination of the above. The system further includes a classifier that determines a distance that the vehicle image descriptor/representation is positioned in the projected feature space relative to a hyper-plane. The classifier determines whether the distance meets a threshold and classifies the image when the threshold is met. A processor implements the modules. A graphic user interface outputs the classification.

Claims

1. A method for detecting a vehicle occupancy violation, the method comprising:

acquiring an image including a vehicle from an associated image capture device positioned to view oncoming traffic;

processing pixels of the image for computing a feature descriptor describing a cabin region of the vehicle;

determining a distance that the feature descriptor is positioned in a projected feature space relative to a hyper-plane;

in response to the distance meeting and exceeding a predetermined threshold, applying the feature descriptor to a classifier for classifying the image into respective classes including at least classes for a candidate violator and a non-violator; and

,

outputting the classification.

2. The method of claim 1 further comprising:

locating a windshield of the vehicle in the captured image; and

,

extracting the windshield region from the image and eliminating other regions.

3. The method of claim 1 further comprising:

locating a windshield of the vehicle in the captured image;

defining a passenger side and a driver side within the windshield region; and

,

extracting the passenger side of the windshield region from the image and eliminating the driver side.

4. The method of claim 1, wherein the processing the image is performed by a process selected from a group consisting of:

a Successive Mean Quantization Transform (SMQT);

a Scale-Invariant Feature Transform (SIFT);

a Histogram of Gradients (HOG);

a Bag-of-Visual-Words Representation;

a Fisher Vector (FV) Representation; and

,

a combination of the above.

5. The method of claim 1 further comprising:

before acquiring the image, training a classifier using at least a first sample image including a driver and a passenger occupying a vehicle and at least a second sample image including a driver and no passenger occupying the vehicle.

6. The method of claim 1 further comprising:

in response to the vehicle being classified as a candidate violator, performing further processing on the image for determining whether the vehicle is one of a violator and non-violator.

7. The method of claim 1 wherein the outputting further includes:

providing notice of a candidate violator to an associated user device.

8. A computer program product comprising non-transitory computer-readable recording media which encodes instructions for performing the method of claim 1.

9. A system for detecting a vehicle occupancy violation, the system comprising:

a determination device comprising memory which stores instructions for performing the method of claim 1 and a processor, in communication with the memory for executing the instructions.

10. A system for detecting a vehicle occupancy violation, the system comprising:

an image capture module adapted to acquire an image including a cabin of a vehicle from an associated image capture device positioned to view oncoming traffic;

a violation determination device adapted for processing the image, the device including;

a feature extraction module adapted to compute from pixels of the image a feature descriptor describing a cabin region of the vehicle,

a classifier adapted to;

determine a distance that the feature descriptor is positioned in a projected feature space relative to a hyper-plane;

in response to the distance meeting and exceeding a predetermined threshold, use the feature descriptor for classifying the image into respective classes including at least classes for a candidate violator and a non-violator,

a processor adapted to implement the modules;

a storage device adapted to store classifications associated with corresponding descriptors; and

,

a graphic user interface adapted to output the classification.

11. The system according to claim 10 further comprising an image capture device having near-infrared capability.

12. The system according to claim 10, wherein the feature extraction module is adapted to perform a process selected from a group consisting of:

a Successive Mean Quantization Transform (SMQT);

a Scale-Invariant Feature Transform (SIFT);

a Histogram of Gradients (HOG);

a Bag-of-Visual-Words Representation;

a Fisher Vector (FV) Representation; and

,

a combination of the above.

13. The system according to claim 10, wherein the feature extraction module is adapted to:

divide the image into sub-images;

process each sub-image to determine low-level features representing each sub-image; and

,

concatenate the low-level features into a feature vector.

14. The system according to claim 10 further comprising a user device adapted to receive as notice the classification indicating that that the vehicle is a candidate violator.

15. The system according to claim 10, wherein the violation determination device is further adapted to:

in response to the vehicle being classified as belonging to a candidate violator, perform further processing on the image for determining whether the vehicle is one of a violator and non-violator.

16. The system according to claim 10, wherein the outputting further includes:

providing notice of a candidate violator to an associated user device.

17. The system according to claim 10 further comprising:

a windshield localization module adapted to extract the windshield region from the image.

18. The system according to claim 10, wherein the classifier is adapted to:

acquire a set of sample images including at least a first image of a vehicle with a passenger and a second image of a vehicle with no passenger;

process the sample images for descriptors;

associate each of the first and second images with the classifications; and

,

transmit the associations to the storage device for storing the classifications.

19. A system for detecting a vehicle occupancy violation, the system comprising:

an image capture module adapted to acquire an image including a cabin of a vehicle from an associated image capture device positioned to view oncoming traffic;

a violation determination device adapted for processing the image, the device including;

a feature extraction module adapted to;

process pixels of the image for determining a feature descriptor of the image, wherein the process is selected from a group consisting;

a Successive Mean Quantization Transform (SMQT);

a Scale-Invariant Feature Transform (SIFT);

a Histogram of Gradients (HOG);

a Bag-of-Visual-Words Representation;

a Fisher Vector (FV) Representation; and

, a combination of the above;

a classifier adapted to;

determine a distance that the feature descriptor is positioned in a projected feature space relative to a hyper-plane;

apply the feature descriptor to a distance based thresholding parameter; and

in response to the distance meeting and exceeding the distance based thresholding parameter, selectively classify the image;

a processor adapted to implement the modules; and

,

a graphic user interface adapted to output the classification.