APPARATUS AND METHOD FOR DETECTING PEDESTRIANS

• US 20150146919A1
• Filed: 04/23/2014
• Published: 05/28/2015
• Est. Priority Date: 11/28/2013
• Status: Active Grant

• Current Assignee: Hyundai Mobis Co., Ltd.
• Original Assignee: Hyundai Mobis Co., Ltd.
• Inventors: KANG, Byung Jun, RYU, Dong Kyu, CHO, Joon Bum
• Granted Patent: US 9,349,043 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 20/588 Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition

Abstract

Provided is an image processing apparatus for detecting pedestrians. The image processing apparatus includes a lane detecting module configured to extract a lane coordinate value from an input image and a pedestrian detecting module configured to set, as a pedestrian region of interest (ROI), a region between a first line passing through ends of first left and right lanes and a second line passing through ends of second left and right lanes which are respectively disposed above the left and right lanes, and search for the pedestrian ROI by using a predetermined window to detect a pedestrian region having a pedestrian feature.

Claims

1. An image processing apparatus for detecting pedestrians, the image processing apparatus comprising:

a lane detecting module configured to extract a lane coordinate value from an input image; and

a pedestrian detecting module configured to set, as a pedestrian region of interest (ROI), a region between a first line passing through ends of first left and right lanes and a second line passing through ends of second left and right lanes which are respectively disposed above the left and right lanes, and search for the pedestrian ROI by using a predetermined window to detect a pedestrian region having a pedestrian feature.

2. The image processing apparatus of claim 1, wherein the pedestrian detecting module calculates a lane interval which is defined as a length of a line connecting ends of the first left and right lanes, and sets the window by using the calculated lane interval.

3. The image processing apparatus of claim 2, wherein a size of the window is proportional to the lane interval.

4. The image processing apparatus of claim 2, wherein the window includes a width, which is calculated by using the lane interval and a ratio variable used to calculate a width of a pedestrian from the lane interval, and a height which is calculated by using the width and a ratio variable between the width and a height of the pedestrian.

5. The image processing apparatus of claim 1, wherein the pedestrian detecting module divides the pedestrian region into an upper body region and a lower body region, and when the pedestrian feature is detected from at least one of the upper body region and the lower body region, the pedestrian detecting module detects the pedestrian region as a final pedestrian region.

6. The image processing apparatus of claim 5, wherein when the pedestrian feature corresponding to an upper body of the pedestrian is detected from the upper body region, and a pedestrian feature corresponding to a lower body of the pedestrian is detected from the lower body region, the pedestrian detecting module detects the pedestrian region as a final pedestrian region.

7. The image processing apparatus of claim 6, wherein the pedestrian detecting module summates a first result value, which is obtained by digitizing the upper body feature detected by using an upper body window, and a second result value which is obtained by digitizing the lower body feature detected by using a lower body window, and when the summation result value is equal to or greater than a threshold value, the pedestrian detecting module detects the pedestrian region as a final pedestrian region.

8. The image processing apparatus of claim 7, wherein the pedestrian detecting module summates the first result value with a first weight value given thereto and the second result value with a second weight value given thereto.

9. The image processing apparatus of claim 1, wherein the pedestrian feature is a histogram of gradient (HOG) feature.

10. A pedestrian detecting method in an image processing apparatus, the pedestrian detecting method comprising:

extracting a lane coordinate value from an input image;

setting a pedestrian region of interest (ROI) and a lane interval in the input image by using the extracted lane coordinate value;

setting a search window by using the set lane interval; and

detecting a pedestrian region in the set pedestrian ROI by using a pedestrian detection algorithm.

11. The pedestrian detecting method of claim 10, wherein the setting of a search window comprises:

setting a width of the search window by using the lane interval and a ratio variable used to calculate a width of a pedestrian from the lane interval; and

setting a height of the search window by using the set width and a ratio variable between the width and a height of the pedestrian.

12. The pedestrian detecting method of claim 10, wherein the detecting of a pedestrian region comprises:

dividing the pedestrian ROI into an upper body region and a lower body region; and

when the pedestrian feature is detected from at least one of the upper body region and the lower body region by using the pedestrian detection algorithm, detecting a corresponding pedestrian candidate region, from which the pedestrian feature is detected, as the pedestrian region.

13. The pedestrian detecting method of claim 10, wherein,

the pedestrian feature comprises an upper body feature and lower body feature of a pedestrian, and

the detecting of a pedestrian region comprises:

dividing the pedestrian ROI into an upper body region and a lower body region;

extracting a histogram of gradient (HOG) feature corresponding to the upper body feature from the upper body region;

extracting an HOG feature corresponding to the lower body feature from the lower body region; and

when the HOG features are respectively extracted from the upper body region and the lower body region, detecting a pedestrian candidate region, including the upper body region and the lower body region, as the pedestrian region.

14. The pedestrian detecting method of claim 13, wherein the detecting of a pedestrian region comprises:

summating a first result value, which is obtained by digitizing a detection result of the upper body feature detected from the upper body region, and a second result value which is obtained by digitizing a detection result of the lower body feature detected from the lower body region;

comparing the summation result value and a threshold value; and

when the summation result value is equal to or greater than the threshold value, detecting the pedestrian candidate region as the pedestrian region.

15. The pedestrian detecting method of claim 14, wherein the summating comprises:

giving a first weight value to the first result value;

giving a second weight value, which is lower than the first weight value, to the second result value; and

summating the first result value with the first weight value given thereto and the second result value with the second weight value given thereto.