Image-based barrier detection and warning system and method thereof

• US 8,913,128 B2
• Filed: 02/23/2011
• Issued: 12/16/2014
• Est. Priority Date: 12/28/2010
• Status: Active Grant

• Current Assignee: Automotive Research & Testing Center
• Original Assignee: Automotive Research & Testing Center
• Inventors: Chen, Chia-Tseng, Su, Yi-Feng, Wu, Yu-Hsiang, Hsieh, Ming-Feng
• Primary Examiner(s): Patel, Jay
• Assistant Examiner(s): MIKESKA, NEIL R
• Application Number:
  US13/033,154
• Publication Number:
  US 20120162415A1
• Time in Patent Office: 1,392 Days
• Field of Search: 348/142, 348/148, 701/41, 382/103
• US Class Current: 348/142
• CPC Class Codes: H04N 7/18 Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Abstract

The present invention discloses an image-based barrier detection and warning system and a method thereof, wherein an image processing technology is used to establish ROI and verify whether a horizontal line signal exists, and wherein the image processing technology determines whether an object is a barrier via detecting contours of an object and detecting distance to the object, and wherein the system timely outputs warning signals and presents distance to a barrier when detecting the barrier. The image-based barrier detection and warning system enables the driver to watch the surroundings of the vehicle directly and clearly and learn the relative positions of barriers, whereby the driver can park more easily and safely.

Claims

1. An image-based barrier detection and warning system comprising:

at least one image capture unit capturing a road image;

at least one processing device electrically connected with said image capture unit, detecting a horizon and defining ROI-H (Region Of Interest-Height) according to a vanishing point proximate the detected horizon to establish a horizontal line signal according to said road image, determining whether said horizontal line signal exists in said ROI-H and whether said horizontal line signal is interrupted by a barrier, calculating a distance between said image capture unit and the barrier in ROI-G (Region Of Interest-Ground), and outputting a barrier warning instruction responsive to detection thereby of horizon interruption by the barrier within an image region defined by said ROI-H and ROI-G; and

at least one warning device receiving said barrier warning instruction from said processing device and emitting at least one barrier warning alert.

2. The image-based barrier detection and warning system according to claim 1, wherein said processing device determines whether to output said barrier warning instruction according to a result of detecting an upper contour and a low contour of a barrier whose horizontal line signal is shaded.

3. The image-based barrier detection and warning system according to claim 1 further comprising an external signal pickup unit electrically connected with said processing device, receiving an external signal and providing said external signal for said processing device, wherein said external signal is a speed signal, a shift signal, or a steering wheel signal.

4. The image-based barrier detection and warning system according to claim 1, wherein said ROI-H of said horizontal line signal is preset by said processing device.

5. The image-based barrier detection and warning system according to claim 1, wherein said ROI-H of said horizontal line signal is determined by at least one traffic lane line in said road image and a virtual lane line.

6. An image-based barrier detection and warning method comprising steps:

(a) capturing a road image;

(b) detecting a horizon and defining ROI-H (Region Of Interest-Height) according to a vanishing point proximate the detected horizon to establish a horizontal line signal according to said road image;

(c) determining whether said horizontal line signal exists in said ROI-H;

(d) determining whether said horizontal line signal is interrupted by shading;

if said horizontal line signal is shaded, proceeding to Step (e);

if said horizontal line signal is not shaded, detecting a contour of an object and determining whether a height and a width of said object are greater than preset thresholds;

if said height and said width of said object are not greater than said preset thresholds, returning to Step (a);

if said height and said width of said object are greater than said preset thresholds, directly undertaking Step (h);

(e) downwardly detecting a horizontal edge from a center of said horizontal line signal in said ROI-H;

(f) determining an upper edge and a lower edge of an object shading said horizontal line signal in said ROI-H and designating said upper edge and said lower edge;

(g) determining whether said object is a barrier according to a position and a height of said horizontal edge;

if said object is not a barrier, returning to Step (a);

if said object is a barrier, undertaking Step (h);

(h) calculating a distance to said barrier;

if said distance is outside a preset distance range in ROI-G (Region Of Interest-Ground), returning to Step (a);

if said distance is within said preset distance range in said ROI-G, undertaking Step (i); and

(i) emitting a warning signal and undertaking Step (a) responsive to detection thereby of horizon interruption by the barrier within an image region defined by said ROI-H and ROI-G.

7. The image-based barrier detection and warning method according to claim 6, wherein said Step (c) further comprises steps:

(c1) setting a detection range of a vanishing horizontal line in said road image;

(c2) undertaking horizontal edge processing on said detection range of said vanishing horizontal line;

(c3) vertically projecting said road image to a horizontal axis;

(c4) undertaking histogram statistics; and

(c5) determining whether a histogram smaller than said preset threshold exists;

if a histogram smaller than said preset threshold exists, returning to Step (b);

if a histogram smaller than said preset thresholds does not exist, undertaking Step (d).

8. The image-based barrier detection and warning method according to claim 6, wherein detecting a contour of an object in said Step (d) further comprises steps:

(d1) detecting edges and detecting an upper half image and a lower half image;

(d2) undertaking horizontal edge calculation and vertical edge calculation and subtracting horizontal edges from vertical edges in said upper half image;

undertaking horizontal edge calculation in said lower half image; and

(d3) using an expansion operation and an erosion operation to enhance definition of said images.

9. The image-based barrier detection and warning method according to claim 6, wherein said ROI-H of the horizontal line signal is preset by a processing device in said Step (b).

10. The image-based barrier detection and warning method according to claim 6, wherein said ROI-H of the horizontal line signal is determined by at least one traffic lane line in said road image and a virtual lane line in said Step (b).