Obstacle recognition method for autonomous robots

• US 10,452,071 B1
• Filed: 02/27/2017
• Issued: 10/22/2019
• Est. Priority Date: 02/29/2016
• Status: Active Grant

• Current Assignee: Ai Incorprated
• Original Assignee: Ai Incorprated
• Inventors: Ebrahimi Afrouzi, Ali, Mehrnia, Soroush
• Primary Examiner(s): Kassa, Yosef
• Application Number:
  US15/442,992
• Time in Patent Office: 967 Days
• Field of Search: 382103, 382153, 382156, 382157, 382159, 382209, 382278, 382291, 700255, 700259, 700261, 901 8
• CPC Class Codes: B25J 9/1676 Avoiding collision or forbidden zones
• CPC Class Codes: B25J 9/1697 Vision controlled systems
• CPC Class Codes: G05D 1/0214 in accordance with safety or protection criteria, e.g. avoiding hazardous areas monitoring the location of vehicles within a certain area, e.g. forbidden or allowed areas, in traffic control systems for road vehicles G08G1/13
• CPC Class Codes: G05D 1/0246 using a video camera in combination with image processing means
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

A method for identifying objects by an autonomous robot. Images of a work environment are captured by an image sensor positioned on the autonomous robot. An object dictionary containing image data regarding a set of objects is preloaded into the system. Captured images are transmitted to an image processor that processes the images to generate feature vector and identify objects within the images by comparison to objects in the object dictionary. Upon identifying an object in an image as an object from the object dictionary, the system executes a set of preprogrammed responses, which may include altering a movement path to avoid colliding with or driving over the object.

Claims

1. A method of identifying objects for autonomous robots comprising:

capturing images of a workspace by an image sensor disposed on an autonomous robot;

transmitting the captured images to a data processing unit disposed on the autonomous robot;

generating a feature vector from the images by the data processing unit;

using the data processing unit to compare objects in the images to objects in a preloaded object dictionary;

identifying a class to which objects belong using an object classification unit; and

,

upon identifying an object from the object dictionary, executing a set of preprogrammed instructions corresponding to the object identified.

2. The method of claim 1 wherein the object dictionary contains any of:

cables, cords, wires, toys, jewelry, undergarments, socks, shoes, shoelaces, feces, keys, food items, plastic bags, earphones, small animals, and tassels.

3. The method of claim 1 wherein preprogrammed instructions include altering a planned navigation path of the autonomous robot to avoid driving over an identified object.

4. The method of claim 1 wherein the preloaded object dictionary is based on a training set in which images of a plurality of examples of the objects in the object dictionary are processed by the data processing unit under varied lighting conditions and camera poses to extract and compile feature vector and characteristics data and associate that feature vector and characteristics data with a corresponding object.

5. The method of claim 4 in which feature vector and characteristics data includes any of:

edge characteristic combinations, basic shape characteristic combinations, and color characteristic combinations.

6. The method of claim 1 in which the image sensor is positioned such that the images capture an area below a main body of the autonomous robot.

7. The method of claim 1 in which, upon identifying an object, feature vector and characteristics data of that object are added to the object dictionary.

8. A method for identifying objects by an autonomous robot comprising:

in a first step, capturing an image of a portion of a workspace by an image sensor disposed on an autonomous robot;

in a second step, transmitting the captured image to a data processing unit disposed on the autonomous robot;

in a third step, generating a feature vector from the image by the data processing unit;

in a fourth step, using the data processing unit to search for objects in the image by comparing the image data to objects in a preloaded object dictionary;

in a fifth step, if one or more objects are found in the image, identifying a class to which each of the objects belongs using an object classification unit; and

,

in a sixth step, upon identifying an object from the object dictionary, executing a set of preprogrammed instructions corresponding to the object identified.

9. The method of claim 8 wherein the object dictionary contains any of:

cables, cords, wires, toys, jewelry, undergarments, socks, shoes, shoelaces, feces, keys, food items, plastic bags, earphones, small animals, and tassels.

10. The method of claim 8 wherein preprogrammed instructions include altering a planned navigation path of the autonomous robot to avoid driving over an identified object.

11. The method of claim 8 wherein the preloaded object dictionary is based on a training set in which images of a plurality of examples of the objects in the object dictionary are processed by the data processing unit under varied lighting conditions and camera poses to extract and compile feature vector and characteristics data and associate that feature vector and characteristics data with a corresponding object.

12. The method of claim 11 in which feature vector and characteristics data includes any of:

edge characteristic combinations, basic shape characteristic combinations, and color characteristic combinations.

13. The method of claim 8 in which the image sensor is positioned such that the images capture an area below a main body of the autonomous robot.

14. The method of claim 8 in which, upon identifying an object, feature vector and characteristics data of that object are added to the object dictionary.