Automated processing method for bus crossing enforcement

• US 9,741,249 B2
• Filed: 08/16/2011
• Issued: 08/22/2017
• Est. Priority Date: 08/16/2011
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Inventors: Burry, Aaron M., Bala, Raja, Fan, Zhigang
• Primary Examiner(s): Geroleo, Francis G
• Application Number:
  US13/210,447
• Publication Number:
  US 20130044219A1
• Time in Patent Office: 2,198 Days
• Field of Search: 348149
• CPC Class Codes: G06V 20/62 Text, e.g. of license plates, overlay texts or captions on TV images
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G08G 1/0175 by photographing vehicles, e.g. when violating traffic rules

Abstract

As set forth herein, systems and methods are described that facilitate to analyze a video stream from a camera mounted on the side of a school bus, wherein a sub-set of video sequences showing cars illegally passing the stopped school bus are automatically identified through image and/or video processing. The described systems and methods provide a significant savings in terms of the amount of manual review that is required to identify such violations. The video sequences also can be analyzed further to additionally produce images of the license plate (for identification of the violator), thereby providing further reduction in required human processing and review time. In one embodiment, automatic license plate recognition (ALPR) is employed to identify text on the violator'"'"'s license plate, as well as the state by which the license plate was issued, without requiring human review of the license plate image.

Claims

1. A computer-implemented method for identifying moving vehicles that illegally pass a school bus during a bus stop, comprising:

receiving a video sequence from a camera device mounted on a school bus;

partitioning the video sequence into video segments such that each video segment corresponds to a single bus stop and comprises one or more video frames captured during the bus stop;

analyzing the frames within each video segment to detect a moving vehicle in one or more of the frames;

identifying and tagging frames in which a moving vehicle is detected;

identifying and tagging video segments that comprise tagged frames;

for each detected moving vehicle, locating a license plate on the moving vehicle;

identifying license plate information comprising the alphanumeric characters on the license plate and the state of origin of the license plate; and

appending metadata, which describes the license plate information, to at least one of the tagged segment and the tagged frame in which the license plate information is identified to generate a violation package;

wherein detecting the moving vehicle comprises;

comparing an initial frame in a video segment to subsequent frames in the video segment to identify differences in pixel intensity;

calculating a sum total error for pixels in each subsequent frame;

comparing the sum total error for each subsequent frame to a predetermined threshold value as a function of the equation;

2. The method according to claim 1, further comprising:

providing the violation package to a law enforcement organization.

3. The method according to claim 1, wherein detecting the moving vehicle comprises:

calculating frame-to-frame pixel intensity differences between each pair of consecutive frames in each video segment;

computing a standard deviation value for pixels in each frame that exhibits a frame-to-frame difference;

comparing the standard deviation values to a predetermined threshold value to identify a frame that include an image of a moving vehicle; and

tagging the identified frame as having an image of a moving vehicle.

4. The method according to claim 3, wherein the frame-to-frame pixel intensity differences between consecutive image frames in the video segments are calculated as a function of the equation:

I_delta(x,y,k)=abs[I_frame(x,y,k)−

I_frame(x,y,k−

1)], k=1,2, . . .

where x and y are the pixel coordinates (row and column) within an image and k is the image frame index, and where I_delta represents the difference in pixel values between frames and I_frame represents a given frame.

5. The method according to claim 1, wherein the camera device records the video sequence periodically by starting to record upon deployment of a stop sign mounted on the school bus and terminating recording upon retraction of the stop sign to the school bus, and wherein partitioning the video sequence into segments is performed as a function of a structure of the video sequence, such that each segment begins at a deployment of the stop sign and ends at a retraction of the stop sign.

6. The method according to claim 1, wherein partitioning the video is performed using timestamp information associated with the video frames and matching the time stamp information to known times of deployment and retraction of a stop sign coupled to the school bus.

7. The method according to claim 1, wherein analyzing the frames further comprises:

obtaining motion vectors for the frames within the video segment;

calculating a sum total of the magnitude of the motion vectors within each frame;

comparing the sum total motion vector magnitude within each frame to a predetermined threshold value;

identifying frames that have a sum total motion vector magnitude greater than the predetermined threshold value as having an image of a moving vehicle.

8. A system that facilitates identifying moving vehicles that illegally pass a school bus during a bus stop, comprising:

a camera device that is mounted to a school bus and that records video whenever a stop sign coupled to the school bus is in a deployed position;

wherein the camera is aimed to capture video of at least a portion of the deployed stops sign and of any moving vehicle that passes the school bus while the stop sign is deployed; and

a processor configured to execute stored computer-executable instructions for;

receiving a video sequence from the camera device;

partitioning the video sequence into video segments such that each video segment corresponds to a single bus stop and comprises one or more video frames captured during the bus stop;

analyzing the frames within each video segment to detect a moving vehicle in one or more of the frames;

identifying and tagging frames in which a moving vehicle is detected;

identifying and tagging video segments that comprise tagged frames;

for each detected moving vehicle, locating a license plate on the moving vehicle;

identifying license plate information comprising the alphanumeric characters on the license plate and the state of origin of the license plate; and

appending metadata, which describes the license plate information, to at least one of the tagged segment data and the tagged frame data in which the license plate information is identified to generate a violation package;

wherein to detect the moving vehicle, the processor executes instructions for;

comparing an initial frame in a video segment to subsequent frames in the video segment to identify differences in pixel intensity;

calculating a sum total error for pixels in each subsequent frame;

comparing the sum total error for each subsequent frame to a predetermined threshold value using the equation;

9. The system according to claim 8, wherein the processor executes instructions for:

transmitting the violation package to a law enforcement organization.

10. The system according to claim 8, wherein to detect the moving vehicle, the processor executes instructions for:

calculating frame-to-frame pixel intensity differences between each pair of consecutive frames in each video segment;

computing a standard deviation value for pixels in each frame that exhibits a frame-to-frame difference;

comparing the standard deviation values to a predetermined threshold value to identify a frame that include an image of a moving vehicle; and

tagging the identified frame as having an image of a moving vehicle.

11. The system according to claim 10, wherein the frame-to-frame pixel intensity differences between consecutive image frames in the video segments are calculated by the processor as a function of the equation:

I_delta(x,y,k)=abs[I_frame(x,y,k)−

I_frame(x,y,k−

1)], k=1,2, . . .

where x and y are the pixel coordinates (row and column) within an image and k is the image frame index, and where I_delta represents the difference in pixel values between frames and I_frame represents a given frame.

12. The system according to claim 8, wherein the camera device records the video sequence periodically by starting to record upon deployment of a stop sign mounted on the school bus and terminating recording upon retraction of the stop sign to the school bus, and wherein partitioning the video sequence into segments is performed as a function of a structure of the video sequence, such that each segment begins at a deployment of the stop sign and ends at a retraction of the stop sign.

13. The system according to claim 8, wherein partitioning the video is performed using timestamp information associated with the video frames and matching the time stamp information to known times of deployment and retraction of a stop sign coupled to the school bus.

14. The system according to claim 8, wherein when analyzing the frames, the processor executes stored computer-executable instructions for:

obtaining motion vectors for the frames within the video segment;

calculating a sum total of the magnitude of the motion vectors within each frame;

comparing the sum total motion vector magnitude within each frame to a predetermined threshold value;

identifying frames that have a sum total motion vector magnitude greater than the predetermined threshold value as having an image of a moving vehicle.

15. A method of identifying moving vehicles that illegally pass a school bus during a bus stop, comprising:

partitioning a video sequence recorded during a bus stop into video segments;

analyzing each frame of the video segment to detect a moving vehicle in one or more of the frames;

tagging a frame in which a moving vehicle is detected;

using automated license plate recognition (ALPR) to identify a license plate number and state of origin of a license plate on the moving vehicle;

appending metadata describing the license plate number and state of origin to the tagged frame data to generate a violation package; and

transmitting the violation package to a law enforcement agency for review;

wherein analyzing each frame of the video further comprises;

comparing an initial frame in a video segment to subsequent frames in the video segment to identify differences in pixel intensity;

calculating a sum total error for pixels in each subsequent frame;

comparing the sum total error for each subsequent frame to a predetermined threshold value as a function of the equation;