Automotive occupant sensing device

• US 5,528,698 A
• Filed: 03/27/1995
• Issued: 06/18/1996
• Est. Priority Date: 03/27/1995
• Status: Expired due to Term

• Current Assignee: Bendix Commercial Vehicle Systems LLC
• Original Assignee: Rockwell International Corporation
• Inventors: Boon, Cathy L., Kamei, John H., Stevens, Phillip F.
• Primary Examiner(s): Couso, Jose L.
• Assistant Examiner(s): Do, Anh Hong
• Time in patent office: 449 Days
• Field of Search: 280/732, 280/735, 250/227.15, 382/100, 382/103, 382/104, 348/143, 348/148
• US Class Current: 382/100
• CPC Class Codes: B60N 2/002 Passenger detection systems
• CPC Class Codes: B60R 21/01538 for image processing, e.g. cameras or sensor arrays
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 2207/30268 Vehicle interior
• CPC Class Codes: G06T 7/0008 checking presence/absence

Abstract

A vehicle occupant safety system includes an image sensor and processor that provides a classification of the occupancy status of the passenger seat area to determine whether an airbag should be deployed. The image processing system includes a photodetector array sensor and lens assembly with image processing electronics to acquire a machine vision representation of the passenger seat area within a vehicle. The objects in the field of view are then discriminated to determine whether a rear-facing child carrier is located in the passenger seat, such that the passenger-side airbag can be disabled.

Claims

1. A vehicle sensor system for sensing the presence of an object in the interior of a vehicle and classifying an image of the object, wherein the object is associated with reflected light energy, comprising:

image classification information representing potential images and features of the object;

storage means for storing the image classification information;

an image sensor for viewing the interior of the vehicle, the image sensor including;

a lens for viewing and focusing the image of the object, and

a photodetector array for acquiring the image of the object, and for detecting the reflected light energy associated with the object, further wherein the reflected light energy is represented as raw analog image signals;

an analog-to-digital converter for receiving the raw analog image signals, and converting the raw analog image signals into digital data representing a low resolution gray-scale outline of the image of the object; and

a processor for analyzing the digital data and for distinguishing the gray-scale outline of the image represented by the digital data, wherein the processor includes

comparison means for comparing the distinguished outline of the image of the object with the stored classification information, wherein if the outline of the image of the object matches one of the stored potential images and features, the processor determines that the outline of the image represents the stored potential image.

2. The vehicle sensor system of claim 1, wherein the photodetector array comprises a plurality of phototransistors.

3. The vehicle sensor system of claim 1, wherein the photodetector array comprises at least one charge coupled device.

4. The vehicle sensor system of claim 1, wherein the vehicle is equipped with a passenger seat carrying a rear-facing child carrier and a passenger-side airbag, further wherein a plurality of potential images of rear-facing child carriers is stored in the storage means, such that if the processor determines that the outline of the image of the object in the passenger seat matches one of the potential images of rear-facing child carders, the processor causes the passenger-side airbag to be disabled.

5. The vehicle sensor system of claim 1, further comprising illumination means for illuminating the interior of the vehicle.

6. The vehicle sensor system of claim 5, wherein the illumination means comprises at least one light emitting diode.

7. The vehicle sensor system of claim 5, wherein the illumination means comprises at least one filtered bulb.

8. The vehicle sensor system of claim 1, wherein the storage means, the image sensor, the analog-to-digital converter, and the processor are coupled together on a circuit board.

9. The vehicle sensor of claim 8, wherein the circuit board is flexible to fit into a variety of configurations.

10. The vehicle sensor of claim 1, wherein the image sensor is mounted in the interior of the vehicle.

11. The vehicle sensor of claim 1, wherein the interior of the vehicle has an overhead console, such that the image sensor is mounted to the overhead console.

12. A safety system for enhancing the safety of a passenger seated in a passenger seat of a vehicle having a passenger-side airbag controlled by airbag electronics, the safety system for disabling the passenger-side airbag if an image representing a child seated in a rear-facing child carrier seat is identified, the system comprising:

a storage device for storing templates of data representing a plurality of seating arrangements in the vehicle, wherein at least one of the templates defines a child seated in a rear-facing child carrier seat;

a sensor for viewing the passenger seat of the vehicle, the sensor including;

a lens for focusing the image of the passenger, the passenger having associated reflected light energy,

a photodetector array for detecting and measuring the reflected light energy of the passenger, the photodetector array providing output gray-scale image signals; and

a processor for comparing the output gray-scale image signals with the stored templates of data to identify whether the output gray-scale image signals correlate with one of the plurality of seating arrangements represented by the stored templates, wherein if a child seated in a rear-facing child carrier seat is identified, the processor communicates with the airbag electronics to disable the passenger-side airbag.

13. The safety system of claim 12, wherein the processor comprises:

means for identifying weak and false edges of the output image signals;

means for segregating the weak and false edges from the output image signals to define an enhanced image;

means for filtering the enhanced image to disregard false edges and weak edges; and

means for modifying the identification of the passenger according to the filtered enhanced image.

14. The safety system of claim 13, further comprising a self-testing system for continuously testing the proper operation of the sensor and processor functions.

15. A method of classifying an image of an object in a vehicle observed by a sensor, the method comprising the steps of:

storing classification data representing conventional configurations of the observed image in a reference library;

acquiring a gray scale input image of the object observed in the vehicle;

distinguishing details and edges of the gray scale input image;

identifying potential false edges of the gray scale input image to differentiate unnatural edges obstructing the gray scale input image;

segregating the false edges from the gray scale input image to define an enhanced image;

determining angles of the edges of the enhanced image to identify high-probability edges and weak edges;

filtering the enhanced image to disregard false edges and weak edges, the filtered enhanced image defining a feature vector space;

correlating the feature vector space with the stored classification data in the reference library; and

classifying the correlated feature vector space according to a match between the feature vector space and the reference library.

16. The method of claim 15, further comprising the steps of:

performing all of the steps of claim 15 multiple times;

storing the classified correlated feature vector space values as history data;

comparing the most recently classified correlated feature vector space with the history data; and

modifying the classified correlated feature vector space according to the comparison with the history data.

17. The method of claim 15, wherein the image observed by the sensor represents an occupant seated in the passenger seat of a vehicle having a passenger-side airbag which is deployed by an airbag electronics system, and further comprising the steps of:

classifying the correlated feature vector space as a child seated in a rear-facing child carrier, the classified correlated feature vector space defining a classified image; and

communicating to the airbag electronics system to disable passenger-side airbag.

18. A method of identifying an image in a passenger seat of a vehicle equipped with a passenger-side airbag, the method comprising the steps of:

storing classification data representing at least three possible seating arrangements of the image, the at least three possible seating arrangements including an empty passenger seat, a child seated in a rear-facing child carrier in the passenger seat, and a forward-facing occupant, the classification data being stored in a reference library;

observing the image, wherein the image is observed for a plurality of flames defined by predetermined periods of time;

acquiring the plurality of frames of the observed image;

interpreting the acquired frames of the image by

distinguishing details and edges of the observed image,

segregating weak and false edges of the observed image, and

filtering the weak and false edges from the observed image;

comparing the filtered observed image with the stored classification data to determine whether the filtered observed image matches one of the at least three possible seating arrangements; and

classifying the filtered observed image within one of the at least three possible seating arrangements, wherein

if the passenger seat is determined to be empty, the passenger-side airbag is disabled,

if a child seated in a rear-facing child carrier is determined to be in the passenger seat, the passenger-side airbag is disabled, and

if a forward-facing occupant is determined to be seated in the passenger seat, the passenger-side airbag is enabled.

19. The method of claim 18, further comprising the steps of:

storing a plurality of previous frames of the observed image as history data;

comparing a most recent classified image with the history data; and

modifying the classified image according to the history data.