Methods and Apparatus for Three-Dimensional (3D) Imaging

• US 20160356890A1
• Filed: 05/05/2016
• Published: 12/08/2016
• Est. Priority Date: 08/26/2014
• Status: Active Grant

• Current Assignee: Massachusetts Institute of Technology
• Original Assignee: Massachusetts Institute of Technology
• Inventors: Frechette, Jonathan, Fried, Dale G.
• Granted Patent: US 9,915,733 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G01S 17/10 using transmission of interrupted, pulse-modulated waves determination of distance by phase measurements G01S17/32
• CPC Class Codes: G01S 17/42 Simultaneous measurement of distance and other co-ordinates indirect measurement G01S17/46
• CPC Class Codes: G01S 17/86 Combinations of lidar systems with systems other than lidar, radar or sonar, e.g. with direction finders
• CPC Class Codes: G01S 17/89 for mapping or imaging
• CPC Class Codes: G01S 17/90 using synthetic aperture techniques
• CPC Class Codes: G01S 7/481 Constructional features, e.g. arrangements of optical elements
• CPC Class Codes: G01S 7/4863 Detector arrays, e.g. charge-transfer gates

Abstract

A method of imaging a scene includes estimating multiple three-dimensional (3D) representations, each of which corresponds to a respective portion of the scene. Neighboring portions of the scene area are at least partially overlapping. Each 3D representation is estimated by illuminating the respective portion of the scene with a light burst including multiple light pulses, after which multiple point clouds are generated by detecting photons reflected or scattered from the respective portion of the scene using a focal plane array. Data points in the point clouds represent a distance between the focal plane array and a scene point in the respective portion of the scene. The 3D representation is then estimated based on the multiple point clouds via coincidence processing. The method then generates a 3D image of the scene based on the multiple 3D representations.

Claims

1. A method of imaging a scene, the method comprising:

A) estimating a first three-dimensional (3D) representation of a first portion of the scene, the estimating of the first 3D representation comprising;

A1) illuminating the first portion of the scene with a first light burst comprising a first plurality of light pulses, the first light burst having a first burst duration equal to or less than 1 millisecond;

A2) generating a first plurality of point clouds, corresponding to the first plurality of the light pulses, by detecting photons reflected or scattered from the first portion of the scene using a focal plane array, a first data point in the first plurality of point clouds representing a first distance between the focal plane array and a first scene point in the first portion of the scene; and

A3) estimating the first 3D representation of the first portion of the scene based at least in part on the first plurality of point clouds;

B) estimating a second 3D representation of a second portion of the scene at least partially overlapping with the first portion of the scene, the estimating of the second 3D representation comprising;

B1) illuminating the second portion of the scene with a second light burst comprising a second plurality of light pulses, the second light burst having a second burst duration equal to or less than 1 millisecond;

B2) generating a second plurality of point clouds, corresponding to the second plurality of the light pulses, by detecting photons reflected or scattered from the second portion of the scene using the focal plane array, a second data point in the second plurality of point clouds representing a second distance between the focal plane array and a second scene point in the second portion of the scene; and

B3) estimating the second 3D representation of the second portion of the scene based at least in part on the second plurality of point clouds; and

C) generating a three-dimensional image of the scene based at least in part on the first 3D representation of the first portion of the scene and the second 3D representation of the second portion of the scene.

2. The method of claim 1, wherein in A1), illuminating the first portion of the scene with the first plurality of light pulses comprises illuminating the first portion of the scene with 2 to 1000 light pulses.

3. The method of claim 1, wherein in A1), illuminating the first portion of the scene with the first plurality of light pulses comprises illuminating the first portion of the scene with 10 to 100 light pulses.

4. The method of claim 1, wherein A1) comprises:

illuminating the first portion of the scene using at least one light pulse having a pulse duration of 0.1 nanosecond to 10 nanoseconds.

5. The method of claim 1, wherein A2) comprises:

detecting the photons reflected or scattered from the first portion of the scene using at least one single-photon detector in the focal plane array.

6. The method of claim 1, wherein A2) comprises:

detecting the photons reflected or scattered from the first portion of the scene using at least one Geiger-mode avalanche photodiode in the focal plane array.

7. The method of claim 1, wherein A2) comprises:

detecting the photons reflected or scattered from the first portion of the scene using the focal plane array comprising at least 16 rows and 16 columns of avalanche photodiodes.

8. The method of claim 1, wherein A3) comprises:

estimating the first 3D representation via coincidence processing of the first plurality of point clouds in an angle-angle range coordinate frame.

9. The method of claim 1, wherein A3) comprises:

assigning a probable range value to the first scene point based at least in part on a first plurality of distances represented by a first plurality of data points in the first plurality of point clouds.

10. The method of claim 1, wherein C) comprises:

aligning the first 3D representation of the first portion of the scene with the second 3D representation of the second portion of the scene using 3D registration.

11. The method of claim 1, wherein the first portion of the scene overlaps with the second portion of the scene by 30% to 60%.

12. The method of claim 1, wherein A2) comprises detecting the photons reflected or scattered from the first portion of the scene using the focal plane array at a first position or a first orientation and B2) comprises detecting the photons reflected or scattered from the second portion of the scene using the focal plane array at a second position or a second orientation.

13. A method of forming a three-dimensional (3D) image of a scene, the method comprising:

(A) illuminating a first part of the scene with a burst of 10 to 1000 pulses of laser radiation, the burst having a duration of less than 1 ms and each pulse having a duration of 0.1 nanosecond to 10 nanoseconds;

(B) detecting photons reflected from the first part of the scene with an array of single-photon-sensitive detectors, the array of single-photon-sensitive detectors having at least 16 rows and 16 columns;

(C) generating a 3D representation of the first part of the scene, the representation of the first part of the scene indicative of a time of arrival and a pixel location for each photon detected in (B);

(D) changing a position and/or orientation of the array of single-photon-sensitive detectors with respect to the scene;

(E) repeating steps (A) through (C) for a second part of the scene so as to generate a 3D representation of the second part of the scene, the second part of the scene at least partially overlapping with the first part of the scene; and

(F) aligning the 3D representation of the first part of the scene with the 3D representation of the second part of the scene so as to form the 3D image of the scene.

14. The method of claim 13, wherein (C) comprises:

estimating a 3D geometry of the first part of the scene based on the 3D representation of the first part of the scene.

15. The method of claim 13, wherein (C) comprises:

removing at least one spurious detection from the 3D representation of the first part of the scene.

16. An apparatus for generating a 3D image of a scene, the apparatus comprising:

a laser to generate bursts of 10 pulses to 1000 pulses of laser radiation for illuminating the scene, each burst having a duration of less than 1 ms and each pulse having a duration of 0.1 nanosecond to 10 nanoseconds;

an array of single-photon-sensitive detectors to detect photons reflected from different parts of the scene, the array of single-photon-sensitive detectors having at least 16 rows and 16 columns, each part of the scene overlapping at least one other part of the scene; and

at least one processor, operably coupled to the array of single-photon-sensitive detectors, to (i) generate 3D representations of the different parts of the scene, each 3D representation indicative of a time of arrival and a pixel location for each photon detected by the array of single-photon-sensitive detectors, and (ii) align the 3D representations of the scene with each other so as to form the 3D image of the scene.