Automated running-engine detection in stationary motor vehicles

• US 9,569,680 B2
• Filed: 02/02/2012
• Issued: 02/14/2017
• Est. Priority Date: 02/02/2012
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Cardoso, George Cunha, Bernal, Edgar A., Pennington, Graham
• Primary Examiner(s): Torrente, Richard
• Assistant Examiner(s): Haghani, Shadan E
• Application Number:
  US13/364,857
• Publication Number:
  US 20130201324A1
• Time in Patent Office: 1,839 Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/10048 Infrared image
• CPC Class Codes: G06T 2207/10141 Special mode during image acquisition
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 7/40 Analysis of texture depth or shape recovery from texture G06T7/529
• CPC Class Codes: G06T 7/44 using image operators, e.g. filters, edge density metrics or local histograms
• CPC Class Codes: G06V 10/145 Illumination specially adapted for pattern recognition, e.g. using gratings

Abstract

A method for idling vehicle detection comprises shining a light source on an exterior surface of a vehicle with an illumination module, collecting light reflected from the exterior surface of the vehicle with a capture module, and processing the collected light reflected from the exterior surface of the vehicle surface with a processing module wherein the processing model determines if the vehicle'"'"'s engine is running or if the vehicle'"'"'s engine is not running.

Claims

1. A method for idling vehicle detection, said method comprising:

shining a low-power, diffuse light source that is transversely coherent and longitudinally coherent, on a surface of a vehicle with an illumination module;

collecting data associated with light reflected from said surface of said vehicle with a capture module; and

processing said collected data associated with said light reflected from said surface of said vehicle surface with a processing module wherein said processing module is configured for;

segmenting said data into a region containing an interference pattern and an unlit region;

identifying a speckle in said region containing an interference pattern associated with said collected light reflected from said surface;

assigning a texture value to each of a plurality of pixels in said speckle;

creating a histogram of said texture values of each of said plurality of pixels;

creating a global texture descriptor associated with said speckle from said histogram; and

comparing said global texture descriptor of said speckle to a reference set of texture descriptors comprising texture descriptors of speckles associated with a vehicle'"'"'s engine running and texture descriptors associated with a vehicle'"'"'s engine not running in order to determine if a temporal characteristic and a spatial characteristic of said speckle is indicative of said vehicle'"'"'s engine running or said vehicle'"'"'s engine not running.

2. The method of claim 1 wherein said capture module comprises at least one of a still camera and a video camera configured to possess a circle of confusion between 2 and 5 pixels thereby allowing light reflected from said surface, of said vehicle to form said speckle pattern comprising an interference pattern on a sensor of at least one of said still camera and said video camera.

3. The method of claim 2 further comprising:

manually training a texture discriminator module to identify speckles associated with a vehicle with a running engine and speckles associated with a vehicle without a running engine.

4. The method of claim 3 wherein said texture appearance of said interference pattern is determined by one of:

a local binary pattern analysis;

a scale-invariant feature transformation; and

a process of Gabor filtering.

5. The method of claim 4, wherein

assigning a texture value to each pixel associated with said speckle further comprises:

assigning each of said pixels a greyscale value;

thresholding pixels in a neighborhood of each of said pixels; and

creating an 8-bit binary number which is converted to said texture value.

6. The method of claim 5, wherein said cc comparison is conducted using one of:

a nearest neighbor classifier;

a support vector machine classifier;

a Bayes classifier; and

a neural network classifier.

7. The method of claim 5, wherein said method further comprises:

shining said light source on a plurality of vehicles distributed in an environment;

manually labeling pixels in a field of view of said capture module known to be stationary;

adopting said labeled pixels as said reference set of texture descriptors; and

monitoring said plurality of vehicles distributed in said environment with said single light source to determine if any of said plurality of vehicles are idling.

8. The method of claim 1 wherein said light source comprises a coherent, low-power light source and wherein said coherent low-power light source is a near infrared light source.

9. A system for detecting idling vehicles comprising:

an illumination module configured to shine a low-power, diffuse light source that is transversely coherent and longitudinally coherent, on a surface of a vehicle;

a capture module configured to collect data associated with light reflected from said surface of said vehicle, said light being propagated by said illumination module; and

a processing module for processing said collected data associated with said light reflected from said surface of said vehicle, wherein said processing model is configured for;

segmenting said data into a region containing an interference pattern and an unlit region;

identifying a speckle in said region containing an interference patter associated with said collected light reflected from said surface;

assigning a texture value to each of a plurality of pixels in said speckle;

creating a histogram of said texture values for each of said plurality of pixels;

creating a global texture descriptor associated with said speckle from said histogram; and

comparing said global texture descriptor of said speckle to a reference set of texture descriptors comprising texture descriptors of speckles associated with a vehicle'"'"'s engine running and texture descriptors associated with a vehicle'"'"'s engine not running in order to determine if a temporal characteristic and a spatial characteristic of said speckle is indicative of aid vehicle'"'"'s engine running or if said vehicle'"'"'s engine not running.

10. The system of claim 9 wherein said capture module comprises at least one of a still camera and a video camera configured to possess a circle of confusion of between 2 and 5 pixels, thereby allowing light reflected from said surface of said vehicle to form said speckle pattern comprising an interference pattern on a sensor of at least one of said still camera and said video camera.

11. The system of claim 10 wherein said processing module is further configured for:

manually training a texture discriminator module to identify speckles associated with a vehicle with a running engine and speckles associated with a vehicle without a running engine.

12. The system of claim 11 wherein said processing module determines a texture appearance of said interference pattern by one of:

a local binary pattern analysis;

a scale-invariant feature transformation; and

a process of Gabor filtering.

13. The system of claim 12 wherein assigning a texture value to each pixel associated with said speckle further comprises:

assigning each of said pixels a greyscale value;

thresholding pixels in a neighborhood of each of said pixels; and

creating an 8-bit binary number which is converted to said texture value.

14. The system of claim 13, wherein said processing module compares said texture appearance metrics according to one of;

a nearest neighbor classifier;

a support vector machine classifier;

a Bayes classifier; and

a neural network classifier.

15. The system of claim 14 wherein said light source comprises a coherent, low-power light source, and wherein said coherent low-power light source is a near infrared light source.

16. The system of claim 15, wherein said system is deployed in an environment and wherein said light source is shone on a plurality of vehicles distributed in said environment and said system:

labels pixels in a field of view of said environment known to be stationary;

adopts said labeled pixels as said reference set of texture descriptors; and

monitors said plurality of vehicles distributed in said environment with said single light source to determine if any of said plurality of vehicles is idling.

17. A method for idling vehicle detection, said method comprising:

shining a low-power, diffuse light source that is transversely coherent and longitudinally coherent, on a surface of a vehicle, wherein said light source comprises a coherent, low-power light source;

collecting data associated with light reflected from said surface of said vehicle with a capture module, wherein said capture module is configured to possess a circle of confusion of between 2 and 5 pixels thereby allowing said image to form an interference pattern on a sensor of said capture module; and

processing said collected data associated with said light reflected from said vehicle surface with a processing module wherein processing module is configured for;

segmenting said data into a region containing an interference pattern and an unlit region;

identifying a speckle in said region containing an interference pattern associated with said collected light reflected from said surface;

assigning a texture value to each of a plurality of pixels in said speckle;

creating a histogram of said texture values for each of said plurality of pixels;

creating a global texture descriptor associated with said speckle from said histogram; and

comparing said global texture descriptor of said speckle to a reference set of texture descriptors comprising texture descriptors of speckles associated with a vehicle'"'"'s engine running and texture descriptors associated with a vehicle'"'"'s engine not running to determine if a temporal characteristic and a spatial characteristic of said speckle is indicative of said vehicle'"'"'s engine running or if said vehicle'"'"'s engine not running.

18. The method of claim 17 wherein said coherent low-power light source is at least one of:

a near infrared light source; and

an infrared light source.

19. The method of claim 18 wherein said texture appearance of said interference pattern is determined by one of:

a local binary pattern analysis;

a scale-invariant feature transformation; and

a process of Gabor filtering.

20. The method of claim 19, further comprising:

shining said light source on a plurality of vehicles distributed in an environment;

manually labeling pixels in a field of view of said capture module known to be stationary;

adopting said labeled pixels as said reference set of texture descriptors; and

monitoring said plurality of vehicles distributed in said environment with said single light source to determine if any of said plurality of vehicles are idling.