System and method for providing 3D imaging

• US 9,303,989 B2
• Filed: 02/22/2011
• Issued: 04/05/2016
• Est. Priority Date: 02/23/2010
• Status: Active Grant

• Current Assignee: Ben Gurion University of The Negev
• Original Assignee: Ben Gurion University of The Negev
• Inventors: Nehmadi, Youval, Guterman, Hugo
• Primary Examiner(s): Harandi, Siamak
• Application Number:
  US13/580,927
• Publication Number:
  US 20120314037A1
• Time in Patent Office: 1,869 Days
• Field of Search: 382/154
• US Class Current: 1/1
• CPC Class Codes: G01C 11/06 by comparison of two or more pictures of the same area

Abstract

The present invention relates to a system and method for providing 3D imaging. The system comprises: a) two or more cameras for allowing to generate stereoscopic information only for selected regions in a common or coincident field of view captured simultaneously by both of said two cameras, in order to provide distance measurements information to object(s) located in said selected regions; b) at least one 3DLR module for providing 3D image, wherein said two cameras and said 3DLR module are positioned in such a way that, at least partially, they will be able to capture similar or coincident field of view information; and c) a processing unit for generating said 3D imaging according to said stereoscopic information and said 3D image by using image processing algorithm(s) in order to provide 3D imaging information in real-time.

Claims

1. A system for providing 3D imaging, comprising:

a) two or more cameras for allowing to generate stereo images from real time stereoscopic data only for selected regions in a common or coincident field of view captured simultaneously by said two or more cameras, in order to provide distance measurements information to object(s) located in said selected regions;

b) at least one 3D Low Resolution (3DLR) module for providing a low resolution 3D image by creating an array of Time of Flight (TOF) measurements in different directions from single points based on calculation of time differences between the emitting of a light pulse and a return signal of the light pulse, wherein said two cameras and said 3DLR module are positioned to at least partially capture similar or coincident field of view information; and

c) a first processing unit for generating a high resolution 3D model by combining said stereoscopic data and said low resolution 3D image, wherein the stereoscopic data is used only on regions in the 3DLR image that the 3DLR module cannot distinguish, thereby reducing the amount of computation time required to extract the 3D information and resulting in accurate 3D imaging information in real-time.

2. A system according to claim 1, further comprising an IR camera for allowing to correlate between the measurements of the 3DLR module with the images provided from the two cameras, wherein the correlation is based on a matching algorithm.

3. A system according to claim 1, in which the 3DLR module is a laser range finder unit configured to provide 3D range image using a laser and a scanning system.

4. A system according to claim 1, in which the 3DLR module comprises a receiver for collecting reflected light, a light emitting unit for directing light beams with different spatial angles to different parts of said receiver and a second processing unit for extracting the light beams and synchronizing them with said receiver, thereby providing a 3DLR module with a TOF module that require non-moving parts.

5. A system according to claim 1, in which the 3DLR module further comprises a polarizer for avoiding crosstalk between vehicles coming from the opposite directions.

6. A system according to claim 1, in which the two cameras and the 3DLR module are located on the same imaginary plane facing the same direction.

7. A system according to claim 1, wherein said two cameras each have a CCD sensor, and said system further comprising optic alignment means for allowing to match pixels between the images provided from the 3DLR module and from the two cameras, by allowing to separate the light into its different wavelengths, so that the wavelengths required for distance measurements will be sent to said 3DLR module and the rest of the wavelengths will be sent to the CCD sensor of each of said two cameras.

8. A system according to claim 7, in which the optic alignment means is a beam splitter.

9. A method for providing 3D imaging, comprising the steps of:

a) capturing images from at least two cameras for allowing to generate stereo images from real time stereoscopic data for selected regions in a common or coincident field of view which is captured simultaneously by said cameras, in order to provide distance measurements information to object(s) located in said selected regions;

b) generating low resolution 3D image data by creating an array of Time of Flight (TOF) measurements in different directions from single points using a 3D Low Resolution (3DLR) module, wherein said two cameras and said 3DLR module are positioned to at least partially capture similar or coincident field of view information; and

c) generating a high resolution 3D model by combining said stereoscopic data with said low resolution 3D image, wherein the stereoscopic data is used only on regions in the 3DLR image that the 3DLR module cannot distinguish, thereby reducing the amount of computation time required to extract the 3D information and resulting in accurate 3D imaging information in real-time.

10. A method according to claim 9, in which the selected regions are determined by correlating the 3D image data only with the corresponding image captured by one of the cameras, in order to identify regions in said 3D image that the 3DLR module is unable to distinguish in a given field of view.

11. A method according to claim 9, in which the stereoscopic data is generated by correlating between the images provided by the two cameras using a real-time matching algorithm.

12. A method according to claim 9, further comprising correlating between the data provided by the 3DLR module with the stereoscopic data is done by an optic alignment means.

13. A method according to claim 9 further comprising the steps of:

a) creating an array or matrix of TOF measurements using the 3DLR module;

b) performing real time stereoscopy by using a matching algorithm thereby providing stereoscopic data; and

c) combining said TOF measurements and said stereoscopic data for creating the high resolution 3D model.