Systems and methods for traffic event detection for vehicles using rolling averages

• US 9,975,482 B2
• Filed: 07/20/2016
• Issued: 05/22/2018
• Est. Priority Date: 02/01/2013
• Status: Active Grant

• Current Assignee: Eric Sinclair
• Original Assignee: Eric Sinclair
• Inventors: Sinclair, Eric
• Primary Examiner(s): Le, Vu
• Assistant Examiner(s): Beg, Samah
• Application Number:
  US15/215,394
• Publication Number:
  US 20160328629A1
• Time in Patent Office: 671 Days
• Field of Search: None
• CPC Class Codes: B60Q 9/008 for anti-collision purposes
• CPC Class Codes: B60R 11/04 Mounting of cameras operative during drive; Arrangement of controls thereof relative to the vehicle
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/30236 Traffic on road, railway or crossing
• CPC Class Codes: G06T 2207/30252 Vehicle exterior; Vicinity of vehicle
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 20/584 of vehicle lights or traffic lights
• CPC Class Codes: G08G 1/166 for active traffic, e.g. moving vehicles, pedestrians, bikes
• CPC Class Codes: H04N 23/54 Mounting of pick-up tubes, electronic image sensors, deviation or focusing coils

Abstract

Platforms and techniques are described in which a traffic event detection system includes a camera connected to a CPU in a vehicle. The camera is mounted on an extendable support, and captures images in front of the vehicle at a height generally above that of vehicles that may be ahead of the subject vehicle. The CPU can analyze those images for sudden change in traffic pattern based on a change in red color content of the image stream, and can warn the driver with an alert if the cars in front of him or her brake suddenly, or other traffic events occur. In implementations, short-term and long-term signal averages can be used to reduce false positive signals, and increase overall accuracy of traffic detection processing. In embodiments, the system can apply a timer to determine if potential hazards continue to be detected, and generated louder alarms or other graduated alerts.

Claims

1. A sensor assembly, comprising:

a sensor, the sensor being adapted to be attached to an extendable support and configured to generate an image stream of an area in front of the vehicle for automated traffic event detection; and

control logic, connected to the sensor, the control logic being configured to;

automatically analyze the image stream, and

detect a traffic event in front of the vehicle based on a comparison of rolling averages generated based on at least two sets of images from the image stream wherein the rolling averages comprise at least a short-term rolling average based on a short-term image set and a long-term rolling average based on a long-term image set, and the long-term rolling average is generated by dividing a sum of red color content for matching pixels in the long-term image set by the number of images contained in the long-term image set, and

wherein the detecting a traffic event comprises detecting whether the short-term rolling average plus a selected threshold exceeds the long-term rolling average and detecting a change in average red-color intensity in the short-term rolling average compared to the long-term rolling average.

2. The assembly of claim 1, wherein the detecting a traffic event comprises detecting whether the short-term rolling average plus a selected threshold exceeds the long-term rolling average for a selected period of time.

3. The assembly of claim 2, further comprising—

generating a first alert to a driver of the vehicle when the short-term rolling average plus a selected threshold exceeds the long-term rolling average, and

generating at least a second alert to a driver of the vehicle when the short-term rolling average plus a selected threshold exceeds the long-term rolling average for the selected period of time.

4. A method of automatically detecting traffic events, comprising:

automatically analyzing an image stream of an area in front of a vehicle generated by at least one sensor, and

detecting a traffic event in front of the vehicle based on a comparison of rolling averages generated based on at least two sets of images from the image stream, wherein the detecting a traffic event comprises detecting a change in average red-color intensity in the short-term rolling average compared to the long-term rolling average, and

wherein the rolling averages comprise at least a short-term rolling average based on a short-term image set and a long-term rolling average based on a long-term image set, and the long-term rolling average is generated by dividing a sum of red color content for matching pixels in the long-term image set by the number of images contained in the long-term image set, and

wherein the detecting a traffic event comprises detecting whether the short-term rolling average plus a selected threshold exceeds the long-term rolling average.

5. The method of claim 4, wherein the detecting a traffic event comprises detecting whether the short-term rolling average plus a selected threshold exceeds the long-term rolling average for a selected period of time.

6. The method of claim 5, further comprising—

generating a first alert to a driver of the vehicle when the short-term rolling average plus a selected threshold exceeds the long-term rolling average, and

generating at least a second alert to a driver of the vehicle when the short-term rolling average plus a selected threshold exceeds the long-term rolling average for the selected period of time.