Method and system for detecting signal color from a moving video platform

1. A method for detecting a color of a signal light from a moving vehicle comprising:

capturing an image from a camera mounted on the moving vehicle;

extracting candidate regions in the image that contain a light region based on a database comprising locations of signal structures along railway tracks;

classifying the candidate regions as a signal light region or a non-signal light region; and

identifying the color of the signal light region;

wherein extracting candidate regions in the image that contain a light region comprises mapping a set of measurements in world coordinates to a set of measurements in image coordinates based on a projection matrix.

2. The method of claim 1 , wherein the moving vehicle comprises a railroad train.

3. The method of claim 1 , wherein extracting the candidate regions comprises calibrating the camera.

4. The method of claim 1 , wherein extracting the candidate regions comprises extracting a location and a size of the candidate region using a filter.

5. The method of claim 4 , wherein the filter comprises a Difference of Gaussians (DoG) filter.

6. The method of claim 1 , wherein classifying the candidate regions comprises extracting the candidate regions of predefined sizes and predefined shapes.

7. The method of claim 1 , wherein identifying the color of the signal light region comprises measuring color pixels in an inner core and an outer border of the signal light region.

8. The method of claim 1 , further comprising tracking the color of the signal light region in the image to eliminate false detections.

9. A system for detecting a color of a signal light comprising:

a camera configured to capture images from a moving vehicle;

a filter configured to extract light regions from candidate regions in the image; and

a classifier configured to classify the light regions as a signal light region or a non-signal light region and to detect the color of the signal light region;

wherein candidate regions are determined by mapping a set of measurements in world coordinates to a set of measurements in image coordinates based on a projection matrix.

10. The system of claim 9 , wherein the camera comprises a video camera or a high-speed digital camera.

11. The system of claim 9 , wherein the camera comprises a pan-tilt-zoom camera.

12. The system of claim 9 , wherein the moving vehicle comprises a railroad train.

13. The system of claim 9 , wherein the candidate region is provided by camera calibration.

14. The system of claim 9 , wherein the filter and the classifier are implemented by appropriate programming of a digital processor.

15. The system of claim 9 , wherein the filter comprises difference of gaussians filter.

16. A non-transitory computer-readable storage medium comprising computer-readable instructions of a computer program that, when executed by a processor, cause the processor to perform a method for detecting a color of a signal light in an image comprising:

extracting candidate regions in the image that contain a light region;

classifying the candidate regions as a signal light region or a non-signal light region;

identifying color of the signal light region;

wherein candidate regions are determined based on a database comprising locations of signal structures along railway tracks; and

wherein extracting candidate regions in the image that contain a light region comprises mapping a set of measurements in world coordinates to a set of measurements in image coordinates based on a projection matrix.

17. The non-transitory computer-readable storage medium of claim 16 comprising a code for detecting a location and a size of the light region.

18. The non-transitory computer-readable storage medium of claim 16 comprising a code for measuring color pixels in an inner core and an outer border of the signal light region.

19. A method for detecting a state of signal lights from a moving vehicle comprising:

capturing a video from a camera mounted on the moving vehicle;

extracting images from the video;

detecting light regions in the image corresponding to the signal light;

identifying color of the signal lights;

analyzing color of the signal lights to obtain a visual cue based on a set of decision rules; and

utilizing a tracking algorithm to detect the state of the signal light to eliminate a false detection;

wherein detecting light regions in the image corresponding to the signal light comprises mapping a set of measurements in world coordinates to a set of measurements in image coordinates based on a projection matrix.

20. The method of claim 19 , wherein the state of the signal light comprises an on state or an off state or a blinking state.

21. The method of claim 19 , wherein analyzing color of the signal lights comprises comparing color of two or more signal lights.

22. The method of claim 19 , wherein analyzing color of the signal lights comprises identifying color of the signal light in two or more sequential images.

23. The method of claim 1 , wherein identifying the color of the signal light region comprises utilizing an image histogram comprising tonal pixels of color variations in the image.