Method of operation for a vision-based occupant classification system

• US 7,308,349 B2
• Filed: 06/06/2005
• Issued: 12/11/2007
• Est. Priority Date: 06/06/2005
• Status: Expired due to Fees

• Current Assignee: Delphi Technologies, Inc. (BorgWarner Incorporated)
• Original Assignee: Delphi Technologies, Inc. (BorgWarner Incorporated)
• Inventors: Kong, Hongzhi, Bailey, John R.
• Primary Examiner(s): ARTHUR JEANGLAUDE, GERTRUDE
• Application Number:
  US11/145,708
• Publication Number:
  US 20060276945A1
• Time in Patent Office: 918 Days
• Field of Search: 701/45, 701/46, 701/47, 180/268, 180/271, 180/272, 180/273, 280/734, 280/735
• US Class Current: 701/45
• CPC Class Codes: B60R 21/01538 for image processing, e.g. cameras or sensor arrays
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition

Abstract

A vision-based occupant classification method utilizes a static classification routine for achieving highly accurate occupant classification, a dynamic detection routine for quickly responding to changes in occupant position, and a category state change routine for detecting potential changes in occupant category. The static classification routine determines an initial classification of the occupant, after which the dynamic detection routine monitors for changes in occupant position and the category state change routine monitors for a potential change in occupant category. If category state change routine detects a potential change in occupant category, the static classification routine is re-executed to re-classify the occupant. The category state change routine identifies a composite optical flow vector associated with occupant movement, and detects a potential change in occupant category when the composite flow vector crosses one or more pre-defined entry/exit motion gates.

Claims

1. A method of operation for vision-based occupant classification in a vehicle cabin, comprising the steps of:

producing a series of digital images of a region of said cabin that includes a vehicle seat;

executing a static classification routine that processes the produced digital images to detect and classify an occupant of said seat;

once said static classification routine has detected and classified the occupant of said seat, suspending further execution of said static classification routine, and instead executing a dynamic detection routine that processes the produced digital images to detect a change in position of the detected occupant;

executing a category state change routine following the suspension of said static classification routine for detecting a potential change in an occupant category of said seat; and

if said potential change in occupant category is detected by said category state change routine, suspending further execution of said dynamic detection routine and said category state change routine, and instead re-executing said static classification routine to re-detect and re-classify the occupant of said seat.

2. The method of operation of claim 1, wherein said static classification routine produces an empty seat classification if said seat is empty and an infant seat classification if said seat is occupied by an infant or child car seat, and said method includes the step of:

inhibiting the execution of said dynamic detection routine if said static classification routine produces said empty seat classification or said infant seat classification.

3. The method of operation of claim 1, where an execution time of said dynamic detection routine is significantly shorter than an execution time of said static classification routine.

4. A method of operation for vision-based occupant classification in a vehicle cabin, comprising the steps of:

producing a series of digital images of a region of said cabin that includes a vehicle seat;

processing the produced digital images to detect and classify an occupant of said seat;

once the occupant of said seat has been detected and classified, processing the produced digital images to detect a change in position of the identified occupant;

detecting a potential change in an occupant category of said seat; and

repeating the step of processing the produced digital images to re-detect and re-classify the occupant of said seat when said potential change in occupant category is detected.

5. The method of operation of claim 4, where the step of detecting a potential change in an occupant category of said seat includes the steps of:

defining at least one motion gate adjacent said seat;

processing a region of the produced digital images that corresponds to said motion gate to identify object movement in said region;

determining a composite movement vector from the identified object movement; and

detecting a potential change in the occupant category when said composite movement vector is indicative of occupant movement to or from said seat.

6. The method of operation of claim 5, including the steps of:

computing optical flow vectors in said region; and

summing said optical flow vectors to produce said composite movement vector.

7. The method of operation of claim 5, including the steps of:

detecting a potential change in the occupant category when a direction of said composite movement vector is in a direction window indicative of movement to or from said seat and a magnitude of said composite movement vector is within a predefined range of magnitudes indicative of a change in occupant category.

8. The method of operation of claim 7, including the steps of:

measuring a duration for which the direction of said composite movement vector is in said direction window and the magnitude of said composite movement vector is within said predefined range of magnitudes; and

detecting a potential change in the occupant category when the measured duration exceeds a reference duration.

9. The method of operation of claim 5, wherein said at least one motion gate is between said seat and a vehicle door adjacent said seat.

10. The method of operation of claim 5, wherein said at least one motion gate is between said seat and an adjacent seat in said cabin.