System and method for estimating daytime visibility

• US 9,398,227 B2
• Filed: 01/05/2012
• Issued: 07/19/2016
• Est. Priority Date: 01/05/2012
• Status: Expired due to Fees

• Current Assignee: GM Global Technology Operations LLC (General Motors Company)
• Original Assignee: GM Global Technology Operations LLC (General Motors Company)
• Inventors: Lerner, Ronen
• Primary Examiner(s): Rao, Andy
• Assistant Examiner(s): EDWARDS, TYLER B
• Application Number:
  US13/343,762
• Publication Number:
  US 20130176425A1
• Time in Patent Office: 1,657 Days
• US Class Current: 1/1
• CPC Class Codes: H04N 21/8153 comprising still images, e.g. texture, background image
• CPC Class Codes: H04N 23/71 Circuitry for evaluating the brightness variation

Abstract

A system for estimating daytime visibility conditions includes a camera configured to capture a video sequence over a period time, an odometry sensor configured to monitor motion of the camera, and a video processing unit in communication with the camera and with the odometry sensor. The video processing unit is configured to receive the video sequence from the camera, detect an object within the video sequence, estimate a distance between the camera and the object for each frame of the video sequence, determine a mean perceived brightness of the object for each image frame, and estimate an ambient visibility using the estimated distance and mean perceived brightness of the object.

Claims

1. A system for estimating daytime visibility conditions comprising:

a camera configured to capture a video sequence over a period of time, the video sequence including a plurality of image frames;

an odometry sensor configured to monitor motion of the camera and output odometry data indicative of the monitored motion;

a video processing unit in communication with the camera and with the odometry sensor, the video processing unit configured to;

receive the video sequence from the camera;

detect an object within the video sequence;

estimate a distance between the camera and the object for each of the plurality of image frames;

determine a mean perceived brightness of the object for each of the plurality of image frames; and

estimate an ambient visibility by;

assembling a plurality of point pairs, each point pair including an estimated distance of the object and a determined mean perceived brightness of the object, and corresponding to the object in a different respective one of the plurality of image frames;

fitting a curve to the plurality of point pairs, the curve based on a range-dependent object brightness model; and

determining a scattering coefficient from the fitted curve, wherein the scattering coefficient is indicative of an estimate of ambient visibility.

2. The system of claim 1, wherein the video processing unit is further configured to compare the estimate of ambient visibility to a threshold, and to perform a control action if the estimate exceeds the threshold.

3. The system of claim 2, further comprising a light source, and wherein the control action includes illuminating the light source.

4. The system of claim 1, wherein detecting an object within the video sequence includes:

detecting an image region having defined boundaries; and

examining the motion of the image region between at least two of the plurality of image frames for motion of the object within the frame.

5. The system of claim 1, wherein the video processing unit is configured to estimate the distance between the camera and the object using the varying location of the object between at least two of the plurality of image frames, together with the received odometry data.

6. The system of claim 5, wherein the video processing unit is further configured to estimate the distance between the camera and the object using a structure-from-motion processing technique.

7. The system of claim 1, wherein the odometry sensor includes a wheel speed sensor and a steering angle sensor.

8. A method of estimating an ambient visibility comprising:

acquiring a video sequence from a camera associated with a moving vehicle, the video sequence including a plurality of image frames;

detecting an object within the video sequence;

estimating a distance between the camera and the object for each of the plurality of image frames;

determining a mean perceived brightness of the object for each of the plurality of image frames; and

estimating an ambient visibility by;

assembling a plurality of point pairs, each point pair including an estimated distance of the object and a determined mean perceived brightness of the object, and corresponding to the object in a different respective one of the plurality of image frames;

fitting a curve to the plurality of point pairs, the curve based on a range-dependent object brightness model; and

determining a scattering coefficient from the fitted curve, wherein the scattering coefficient is indicative of an estimate of ambient visibility.

9. The method of claim 8, further comprising:

comparing the estimate of ambient visibility to a threshold; and

performing a control action if the estimate exceeds the threshold.

10. The method of claim 9, wherein performing a control action includes illuminating a light source.

11. The method of claim 9, wherein performing a control action includes speed-limiting an engine.

12. The method of claim 8, wherein detecting an object within the video sequence includes:

detecting an image region having defined boundaries; and

examining the motion of the image region between at least two of the plurality of image frames for motion of the object within the frame.

13. The method of claim 8, further comprising receiving data indicative of the motion of the camera;

and

wherein estimating the distance between the camera and the object includes analyzing differences in the appearance of the object between at least two of the plurality of image frames, in view of the received motion data.

14. The method of claim 8, wherein the object is a first object;

and further comprising;

detecting a second object within the video sequence;

estimating a distance between the camera and the second object for each of the plurality of image frames;

determining a mean perceived brightness of the second object for each of the plurality of image frames; and

wherein estimating an ambient visibility includes determining a scattering coefficient from the estimated distances and mean perceived brightnesses of the first and second objects across the plurality of image frames.

15. The method of claim 8, further comprising fusing perceived radar information with the acquired video sequence.

16. The method of claim 8, further comprising adjusting the weighting of one or more perception systems in a composite sensor fusion using the estimate of ambient visibility.