METHOD AND DEVICE FOR COMPUTING A CHANGE IN AN IMAGE SCALE OF AN OBJECT

• US 20150170390A1
• Filed: 07/10/2013
• Published: 06/18/2015
• Est. Priority Date: 07/10/2012
• Status: Active Grant

• Current Assignee: Robert Bosch GmbH
• Original Assignee: Robert Bosch GmbH
• Inventors: Conomis, Christos
• Granted Patent: US 9,384,576 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G01S 11/12 using electromagnetic waves other than radio waves
• CPC Class Codes: G06T 11/60 Editing figures and text; Combining figures or text
• CPC Class Codes: G06T 2207/30261 Obstacle
• CPC Class Codes: G06T 3/40 Scaling of whole images or parts thereof, e.g. expanding or contracting
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• 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 is provided for computing a change in an image scale of an object. The method includes a step of recognizing, a step of providing, a step of scaling, a step of ascertaining, and a step of combining. Image coordinates of characteristic image areas are recognized in an instantaneous image in the step of recognizing. The characteristic image areas depict object features at least of the one object in the instantaneous image. A model is provided in the step of providing, model coordinates of the model representing a selection of positions of recognized object features of the object in a preceding image. The model is scaled in the step of scaling in order to obtain a scaled model. At least a portion of the model coordinates is correlated with the recognized image coordinates.

Claims

1-11. -11. (canceled)

12. A method for computing a change in an image scale of an object, comprising:

recognizing image coordinates of characteristic image areas in an instantaneous image, the characteristic image areas depicting object features at least of the one object in the instantaneous image;

providing a model, wherein model coordinates of the model represent a selection of positions of recognized object features of the object in a preceding image;

scaling the model in order to obtain a scaled model, at least a portion of the model coordinates being correlated with the image coordinates recognized in the step of recognizing, and the model and the scaled model differing from one another by a model scaling factor;

ascertaining an image scaling factor using the image coordinates of at least two recognized characteristic image areas, and image coordinates of the two image areas recognized in the preceding image; and

combining the model scaling factor with the image scaling factor in order to obtain the change in the image scale.

13. The method as recited in claim 12, further comprising:

creating the model, wherein chronologically preceding recognized image coordinates from characteristic image areas of the preceding image are used to convert at least the selection of the image coordinates into model coordinates, the image coordinates being based on an image coordinate origin, and the model coordinates being based on a model coordinate origin, the object being at least partially depicted in the selection.

14. The method as recited in claim 13, wherein the selection has a predefined starting dimension in the step of creating.

15. The method as recited in claim 13, wherein the model coordinates are normalized in the step of creating, a minimum value of the model coordinates being zero and a maximum value of the model coordinates being one.

16. The method as recited in claim 12, wherein a dimension of the model and a dimension of the scaled model are detected in the step of scaling, and the model scaling factor is ascertained using a change in the dimension.

17. The method as recited in claim 12, wherein the model scaling factor is weighted with a first weighting factor, and the image scaling factor is weighted with a second weighting factor, in the step of combining.

18. The method as recited in claim 12, wherein the model is at least one of shifted by a translation vector and rotated by a rotation vector in the step of scaling.

19. The method as recited in claim 17, wherein the steps of the method are repeated to compute a further change in the image scale, a chronologically subsequent image depicting the characteristic object features at least of the one object at a later point in time, the scaled model being provided in the step of providing, and being scaled, using the already computed change in the image scale, to create an estimation model, the estimation model being scaled in the step of scaling, using a correction factor, to adapt the estimation model to the image coordinates of the chronologically subsequent image recognized in the step of recognizing, the model scaling factor being ascertained using the computed change and the correction factor.

20. The method as recited in claim 19, wherein the first weighting factor is ascertained in the step of combining, based on the correction factor.

21. A device for computing a change in an image scale of an object, comprising:

an arrangement for recognizing image coordinates of characteristic image areas in an instantaneous image, the characteristic image areas depicting object features at least of the one object in the instantaneous image;

an arrangement for providing a model, wherein model coordinates of the model represent a selection of positions of recognized object features of the object in a preceding image;

an arrangement for scaling the model in order to obtain a scaled model, at least a portion of the model coordinates being correlated with the image coordinates recognized in the step of recognizing, and the model and the scaled model differing from one another by a model scaling factor;

an arrangement for ascertaining an image scaling factor using the image coordinates of at least two recognized characteristic image areas, and image coordinates of the two image areas recognized in the preceding image; and

an arrangement for combining the model scaling factor with the image scaling factor in order to obtain the change in the image scale.

22. A computer program product including program code for carrying out a method for computing a change in an image scale of an object, the method comprising:

recognizing image coordinates of characteristic image areas in an instantaneous image, the characteristic image areas depicting object features at least of the one object in the instantaneous image;

providing a model, wherein model coordinates of the model represent a selection of positions of recognized object features of the object in a preceding image;

scaling the model in order to obtain a scaled model, at least a portion of the model coordinates being correlated with the image coordinates recognized in the step of recognizing, and the model and the scaled model differing from one another by a model scaling factor;

ascertaining an image scaling factor using the image coordinates of at least two recognized characteristic image areas, and image coordinates of the two image areas recognized in the preceding image; and

combining the model scaling factor with the image scaling factor in order to obtain the change in the image scale.