Method and system for monitoring driver

• US 8,587,440 B2
• Filed: 02/11/2010
• Issued: 11/19/2013
• Est. Priority Date: 09/22/2009
• Status: Active Grant

• Current Assignee: Automotive Research & Testing Center
• Original Assignee: Automotive Research & Testing Center
• Inventors: Liao, Yu-Sheng, Weng, Ming-Cong, Chen, Yu-Song
• Primary Examiner(s): Zimmerman, Brian
• Assistant Examiner(s): EUSTAQUIO, CAL J
• Application Number:
  US12/704,142
• Publication Number:
  US 20110068934A1
• Time in Patent Office: 1,377 Days
• Field of Search: 348/143
• US Class Current: 340/575
• CPC Class Codes: A61B 5/18 for vehicle drivers or machine operators
• CPC Class Codes: B60K 28/066 actuating a signalling device B60K28/063 takes precedence
• CPC Class Codes: G06V 20/59 inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
• CPC Class Codes: G06V 40/165 using facial parts and geometric relationships
• CPC Class Codes: G08B 21/06 indicating a condition of sleep, e.g. anti-dozing alarms

Abstract

A method and system for monitoring a driver is disclosed. Firstly, an inner cabin image of a vehicle'"'"'s cabin is continuously captured. Next, a face detection of a driver for the inner cabin image is performed to obtain a face detection result. Next, the inner cabin image and the face detection result corresponding to the inner cabin image are stored. When the system observes that either the face of the driver is not detected or the face is not shown within a reasonable face region of at least two continuous inner cabin images according to each face detection result, the system outputs a warning signal to alert the driver. The invention determines a driving state of a driver in a cabin with a method of analyzing a position of a face or a head and using an image extractor having large capture range and a low resolution.

Claims

1. A method for monitoring a driver comprising steps of:

Step (A);

capturing an inner cabin image of a cabin;

Step (B);

performing a face detection of the driver on said inner cabin image to obtain a face detection result;

Step (C);

storing said inner cabin image and said face detection result corresponding to said inner cabin image;

Step (D);

repeating said above-mentioned steps, and when an accumulated number of stored said inner cabin images reaches a predetermined value, executing steps of;

Step (E);

capturing said inner cabin images at a plurality of adjacent storing time points and corresponding said face detection results thereof;

Step (F);

determining whether at least one corresponding said inner cabin image catches a face of said driver according to each captured said face detection result;

if no, outputting a warning signal to alert said driver; and

if yes, executing a step of;

Step (G);

determining whether said faces of two continuous and corresponding said inner cabin images are shown within a reasonable face region according to each captured said face detection result to define a face region correction process;

said face region correction process including taking a positional coordinate of the face, taking said positional coordinate having a highest existing probability, and deciding a position of said reasonable face region; and

,

(1) if said faces of two continuous and corresponding said inner cabin images are within said reasonable face region, then performing a nose detection of said driver on presently-stored said inner cabin image;

obtaining a positional coordinate of a nose of said driver;

comparing said positional coordinate of said nose with said reasonable face region to obtain a head deflection angle of said driver to define a nose detection process;

determining whether said head deflection angle is lower than a predetermined threshold, and returning to said Step (E); and

(2) if said faces of two continuous and corresponding said inner cabin images are not within said reasonable face region, then outputting said warning signal to alert said driver.

2. The method for monitoring a driver according to claim 1, wherein in said step of performing said face detection of said driver on said inner cabin image, an Adaboost algorithm is used to perform said face detection of said driver on said inner cabin image.

3. The method for monitoring a driver according to claim 1, wherein said reasonable face region is a predetermined region.

4. The method for monitoring a driver according to claim 1, wherein said reasonable face region is a squarely-facing face region.

5. The method for monitoring a driver according to claim 2, wherein said face detection result comprises a catch result and a positional coordinate of said face.

6. The method for monitoring a driver according to claim 5, wherein when a determination result of said Step (F) is yes and before said Step (G), said method further comprises steps of:

taking out captured said inner cabin images catching said face and corresponding said positional coordinate of said face thereof;

analyzing said positional coordinate of said face of each captured said inner cabin image and taking said positional coordinate having a highest existing probability from said positional coordinate of said face of each captured said inner cabin image as determined by said Adaboost algorithm; and

deciding a position of said reasonable face region according to said positional coordinate having said highest existing probability.

7. The method for monitoring a driver according to claim 1, wherein said nose of said driver is detected by an Adaboost algorithm.

8. The method for monitoring a driver according to claim 1, wherein when said Step (E) is executed at two adjacent times, a part of captured said inner cabin images is equal.

17. The method for monitoring a driver according to claim 1, wherein said face detection result comprises a catch result and said positional coordinate of said face.

9. A system for monitoring a driver comprising:

an image extractor for continuously capturing an inner cabin image;

an image signal process unit coupled to said image extractor for performing a face detection of the driver on said inner cabin image to obtain a face detection result;

a control unit coupled to said image signal process unit and a memory, controlling to store said inner cabin image and corresponding said face detection result thereof in said memory, wherein when said control unit observes that a face is not shown in one said inner cabin image or said face is not shown within a reasonable face region of at least two continuous said inner cabin images according to each said face detection result stored in said memory,

said control unit outputs a control signal, said control unit compares said positional coordinate of said nose with said reasonable face region to obtain a head deflection angle of said driver, and wherein when said control unit determines that a face is shown in one said inner cabin image according to said face detection result, said control unit takes a positional coordinate of the face and then takes said positional coordinate having a highest existing probability to decide a position of said reasonable face region; and

an alerter for receiving said control signal and then outputting a warning signal to alert said driver said control unit controls said alerter to output said warning signal when said head deflection angle is higher than or equal to a predetermined threshold.

10. The system for monitoring a driver according to claim 9, wherein said image signal process unit performs said face detection of said driver on said inner cabin image by an Adaboost algorithm.

11. The system for monitoring a driver according to claim 9, wherein said reasonable face region is a predetermined region.

12. The system for monitoring a driver according to claim 9, wherein said reasonable face region is a squarely-facing face region.

13. The system for monitoring a driver according to claim 9, wherein said face detection result comprises a catch result and said positional coordinate of said face.

14. The system for monitoring a driver according to claim 13, wherein said control unit takes said inner cabin images catching said face from said memory according to each said face detection result stored in said memory, and said positional coordinate having said highest existing probability being determined by said Adaboost algorithm.

15. The system for monitoring a driver according to claim 9, wherein said control unit captures said inner cabin images and corresponding said face detection results thereof at a plurality of adjacent storing time points from said memory, and when said control unit observes that said face is not shown in one said inner cabin image or said face is not shown within said reasonable face region of at least two continuous said inner cabin images according to each captured said face detection result, said control unit outputs said control signal.

16. The system for monitoring a driver according to claim 9, wherein said nose of said driver is detected by an Adaboost algorithm.

18. The system for monitoring a driver according to claim 9, wherein said face detection result comprises a catch result and said positional coordinate of said face.