OBJECT ESTIMATION APPARATUS AND OBJECT ESTIMATION METHOD

• US 20160203376A1
• Filed: 08/20/2014
• Published: 07/14/2016
• Est. Priority Date: 08/21/2013
• Status: Active Grant

• Current Assignee: DENSO Corporation, Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: DENSO Corporation, Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Inventors: Ishigami, Hirotake, Shirai, Noriaki, Matsubara, Toshiyuki
• Granted Patent: US 9,767,367 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: B60W 30/06 Automatic manoeuvring for parking
• CPC Class Codes: B60W 30/08 Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06T 2200/04 involving 3D image data
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 2207/30252 Vehicle exterior; Vicinity of vehicle
• CPC Class Codes: G06T 7/269 using gradient-based methods
• CPC Class Codes: G06T 7/292 Multi-camera tracking
• CPC Class Codes: G06T 7/593 from stereo images
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• 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
• CPC Class Codes: G06V 20/64 Three-dimensional objects
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16
• CPC Class Codes: G08G 1/166 for active traffic, e.g. moving vehicles, pedestrians, bikes

Abstract

An object estimation apparatus estimates a position and a speed of an object in images based on the images taken by an image taking means from different positions. The object estimation apparatus includes a shift information taking means that takes temporal shift information of a position of a corresponding pixel from frames arranged in time series in a reference image, which is a reference between the images, a parallax information taking means that takes parallax information of each corresponding pixel from the images with reference to the reference image, an estimated value taking means that estimates estimated values of the position and the speed of the object in three-dimensional space by using a filter based on the shift information taken by the shift information taking means and the parallax information taken by the parallax information taking means, a determination means that determines whether or not each of the shift information taken by the shift information taking means and the parallax information taken by the parallax information taking means is an abnormal value, and a correction means that corrects the estimated value taken by the estimated value taking means based on a determination result of the determination means. The correction means corrects the estimated value taken by the estimated value taking means by using different methods between a case where the determination means determines that the shift information taken by the shift information taking means is an abnormal value and a case where the determination means determines that the parallax information taken by the parallax information taking means is an abnormal value.

Claims

1. An object estimation apparatus, which estimates a position and a speed of an object in images based on the images taken by an image taking means (12) from different positions, characterized in that the object estimation apparatus comprises:

a shift information taking means (28) that takes temporal shift information of a position of a corresponding pixel from frames arranged in time series in a reference image, which is a reference between the images;

a parallax information taking means (30) that takes parallax information of each corresponding pixel from the images with reference to the reference image;

an estimated value taking means (32) that estimates estimated values of the position and the speed of the object in three-dimensional space by using a filter based on the shift information taken by the shift information taking means and the parallax information taken by the parallax information taking means;

a determination means (34) that determines whether or not each of the shift information taken by the shift information taking means and the parallax information taken by the parallax information taking means is an abnormal value; and

a correction means (36) that corrects the estimated value taken by the estimated value taking means based on a determination result of the determination means, and

the correction means corrects the estimated value taken by the estimated value taking means by using different methods between a case where the determination means determines that the shift information taken by the shift information taking means is an abnormal value and a case where the determination means determines that the parallax information taken by the parallax information taking means is an abnormal value.

2. The object estimation apparatus according to claim 1, wherein

when the determination means determines that the shift information taken by the shift information taking means is an abnormal value, the correction means voids the estimated value taken by the estimated value taking means and resets the filter to an initial state.

3. The object estimation apparatus according to claim 1 or 2, wherein

the estimated value taking means takes predicted values of the position and the speed of the object from the estimated value obtained previously, assuming that the object follows a preset process model, and

when the determination means determines that the parallax information taken by the parallax information taking means is an abnormal value, the correction means corrects the estimated value taken by the estimated value taking means to the predicted value taken by the estimated value taking means.

4. The object estimation apparatus according to claim 3, wherein

the determination means determines whether or not the shift information taken by the shift information taking means is an abnormal value based on a difference between the shift information and predicted shift information, which is calculated from the predicted value taken by the estimated value taking means and corresponds to the shift information.

5. The object estimation apparatus according to claim 3 or 4, wherein

the determination means determines whether or not the parallax information taken by the parallax information taking means is an abnormal value based on a difference between the parallax information and predicted parallax information, which is calculated from the predicted value taken by the estimated value taking means and corresponds to the parallax information.

6. An object estimation method, which estimates a position and a speed of an object in images based on the images taken by an image taking means (12) from different positions, characterized in that the object estimation apparatus comprises:

a step (S18) in which a shift information taking means (28) takes temporal shift information of a position of a corresponding pixel from frames arranged in time series in a reference image, which is a reference between the images;

a step (S20) in which a parallax information taking means (30) takes parallax information of each corresponding pixel from the images with reference to the reference image;

a step (S24) in which an estimated value taking means (32) estimates estimated values of the position and the speed of the object in three-dimensional space by using a filter based on the shift information taken by the shift information taking means and the parallax information taken by the parallax information taking means;

a step (S26, S28) in which a determination means (34) determines whether or not each of the shift information taken by the shift information taking means and the parallax information taken by the parallax information taking means is an abnormal value; and

a step (S32, S34, S36) in which a correction means (36) corrects the estimated value taken by the estimated value taking means based on a determination result of the determination means, and

the correction means corrects the estimated value taken by the estimated value taking means by using different methods between a case where the determination means determines that the shift information taken by the shift information taking means is an abnormal and a case where the determination means determines that the parallax information taken by the parallax information taking means is an abnormal.

7. The object estimation method according to claim 6, wherein

when the determination means determines that the shift information taken by the shift information taking means is an abnormal value, the correction means voids the estimated value taken by the estimated value taking means (S34) and resets the filter to an initial state (S32).

8. The object estimation method according to claim 6 or 7, wherein

the estimated value taking means takes predicted values of the position and the speed of the object from the estimated value obtained previously, assuming that the object follows a preset process model, and

when the determination means determines that the parallax information taken by the parallax information taking means is an abnormal value, the correction means corrects the estimated value taken by the estimated value taking means to the predicted value taken by the estimated value taking means (S36).

9. The object estimation method according to claim 8, wherein

the determination means determines whether or not the shift information taken by the shift information taking means is an abnormal value based on a difference between the shift information and predicted shift information, which is calculated from the predicted value taken by the estimated value taking means and corresponds to the shift information.

10. The object estimation method according to claim 8 or 9, wherein

the determination means determines whether or not the parallax information taken by the parallax information taking means is an abnormal value based on a difference between the parallax information and predicted parallax information, which is calculated from the predicted value taken by the estimated value taking means and corresponds to the parallax information.