Image processing apparatus and processing method

• US 4,799,267 A
• Filed: 12/24/1986
• Issued: 01/17/1989
• Est. Priority Date: 10/22/1982
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Yamamoto, Hiroshi, Nakano, Yoshiyuki, Kamejima, Kohji, Iwamoto, Taro, Fujie, Masakatsu
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Couso, Jose L.
• Application Number:
  US06/946,007
• Time in Patent Office: 755 Days
• Field of Search: 318/570, 318/574, 318/580, 318/587, 244/3.17, 358/103, 382/1, 382/15, 382/16, 382/22, 382/25, 382/28, 382/30, 382/31, 382/34, 382/37-40, 901/1, 901/47, 364/449, 364/456, 364/457, 364/460
• US Class Current: 382/153
• CPC Class Codes: B25J 19/023 including video camera means
• CPC Class Codes: G06T 15/10 Geometric effects
• CPC Class Codes: G09B 9/05 the view from a vehicle being simulated G09B9/052, G09B9/058 take precedence

Abstract

In processing topographical information necessary for the traveling of a robot or the like furnished with an automatic traveling function; topography is predicted from the topographical information previously given, so as to prepare a predictive image, a difference image between the predictive image and an actual image caught by a television camera installed on the robot or the like is obtained, and the obtained result if used for correcting the topographical information previously given.

Claims

1. In an apparatus which processes three dimensional topographical information on the arrangement of roads, buildings, obstacles and the like, which are not necessarily located perpendicularly to a direction of movement of an automatic traveling object, for controlling movement of an automatic traveling object, an image processing apparatus comprising:

image input means mounted on the automatic traveling object for viewing a surrounding landscape and providing an output signal representing a two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like;

storage means for storing three dimensional map data as to the arrangement of the roads, building, obstacles and the like in an area where the automatic traveling object is to move;

predicting means connected to said storage means for predicting an output of said image input means for a given location of the automatic traveling object by producing a signal representing a predictive two dimensional intensity distribution of light reflected or emitted from roads, building, obstacles and the like on the basis of said three dimensional map data;

means for comparing an output signal representing an actual two dimensional input intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like from said image input means and the signal representing the predictive two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like and for obtaining a difference two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like; and

correction means connected to said comparing means for updating map data in said storage means on the basis of said difference two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like.

2. An image processing apparatus as defined in claim 1, wherein said predicting means comprises a computer which simulates an optical path that extends from a road, building, obstacle and the like stored in said storage means to generate the two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like and supplies a signal representative thereof to said comparing means.

3. A method which processes three dimensional topographical information on the arrangement of roads, buildings, obstacles and the like, which are not necessarily located perpendicularly to a direction of movement of an automatic traveling object, for controlling the movement of an automatic traveling object, and which employs storage means for storing three dimensional map data and image input means mounted on the automatic traveling object for producing a signal representing a two dimensional intensity distribution of light reflected or emitted from a surrounding landscape viewed from said traveling object, comprising reading out the three dimensional map data from said storage means;

predicting an output of said image input means for a location of the automatic traveling object by calculating a predictive two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like on the basis of said three dimensional map data, comparing said predictive two dimensional intensity distribution to an input represented by a signal delivered from said image input means for obtaining a two dimensional difference intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like; and

updating the three dimensional map data stored in said storage means on the basis of said two dimensional difference intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like.

4. An image processing method as defined in claim 3, wherein at the predicting step, an object represented by map data is retrieved from an area within a viewing angle of a position corresponding to said image input means.

5. An image processing method as defined in claim 3, wherein at the predicting step, a predictive image of only that part of each image within the viewing angle of said image input means is calculated.

6. An image processing method as defined in claim 3, wherein, when a difference has been detected between an input two dimensional intensity distribution of light reflected or emitted from roads, buildings obstacles and the like and a predictive two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like at the comparing step, data relating to the largest object is renewed in said three dimensional map data in a range not contradictory to building, obstacle and the like data in the input two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like and the map three dimensional data, whereupon the three dimensional map data is compared with sequentially received input two dimensional intensity distributions of light reflected or emitted from roads, buildings, obstacles and the like so as to delete any excess part of renewed object data at the updating step.

7. An image processing method as defined in claim 3, wherein, when a difference has been detected between input two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like and the predictive two dimensional intensity distribution of light reflected or emitted from roads, buildings, obstacles and the like at the comparing step, data relating to the smallest object which causes such differences are successively added to the three dimensional map data at the updating step.

8. An image processing method as defined in claim 3, wherein at the updating step, positions where objects causing differences between input two dimensional intensity distributions of light reflected or emitted from roads, buildings, obstacles and the like and predictive two dimensional intensity distributions of light reflected or emitted from roads, buildings, obstacles and the like are classified, and only data relating to objects necessary for determining a traveling scheme of the automatic traveling object are selectively added to the three dimensional map data.

9. An image processing method as defined in claim 3, wherein at the updating step, the positions of the two dimensional difference intensity distributions of light reflected or emitted from roads, buildings, obstacles and the like between input two dimensional intensity distributions of light reflected or emitted from roads, buildings, obstacles and the like and predictive two dimensional intensity distributions of light reflected or emitted from roads, buildings, obstacles and the like are classified, and only data relating to objects necessary for determining a traveling scheme of the automatic traveling object are added to the three dimensional map data.