3D POINT CLOUDS

• US 20170046845A1
• Filed: 04/24/2015
• Published: 02/16/2017
• Est. Priority Date: 04/24/2014
• Status: Active Grant

• Current Assignee: Cathx Research Ltd.
• Original Assignee: Cathx Research Ltd.
• Inventors: Flynn, Michael, Boyle, Adrian
• Granted Patent: US 10,163,213 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 7/521 from laser ranging, e.g. using interferometry; from the projection of structured light
• CPC Class Codes: G06T 7/73 using feature-based methods

Abstract

Provided is a method for generating a 3D point cloud and colour visualisation of an underwater scene, the point cloud comprising a set of (x, y, z) coordinates relating to points in the scene, the method operating in a system comprising at least one camera module, at least one structured light source, and a processing module, each of the at least one camera module being directed at the scene and having substantially the same overlapped field of view.

Claims

1. A method for generating a 3D point cloud and colour visualisation of an underwater scene, the point cloud comprising a set of (x, y, z) coordinates relating to points in the scene, the method operating in a system comprising at least one camera module, at least one structured light source, and a processing module, each of the at least one camera module being directed at the scene and having substantially the same overlapped field of view;

the method comprising;

(a) the at least one structured light source projecting a two-dimensional array of points onto the scene;

(b) each of the at least one camera module capturing an image of the projected array in the scene from first and second positions, such that there is a first point image and second point image;

(c) analysing the images to identify the location of each projected point within each point image;

(d) for each projected point in the captured point images, correlating the location of the point between the first and second point images to calculate the distance to the point from the at least one camera module;

(e) storing the distance and location of the point;

(f) adjusting the position of the two-dimensional array of points in the scene; and

repeating (a) to (e), thereby forming a 3D point cloud;

2. A method as claimed in claim 1, wherein the system further comprises a white light source, and further comprising:

projecting the white light on to the scene in sequence following a structured light acquisition;

each camera acquiring a white light image; and

combining the 3D point cloud and white light images to provide a 3D colour visualisation of the scene.

3. A method as claimed in any preceding claim 1 wherein the system further comprises a plurality of light sources controllable to provide a plurality of illumination profiles;

the method comprising;

illuminating the scene according to a while light profile;

each camera module simultaneously capturing an image of the illuminated scene, such that there is a first scene image and second scene image;

using machine vision to analyse the images to identify at least one feature within each scene image;

comparing features between scene images to identify features that appear in both scene images;

identifying the locations of points in features that appear in both scene images;

correlating the location of the features points between the first and second scene images to calculate the distance to the feature points from the at least one camera module; and

storing the distance and location of the feature points.

4. A method as claimed in claim 1, wherein the at least one camera module comprises a single camera, the method comprising positioning the single camera at a first location, acquiring images and position data at the first location, followed by moving the single camera to a second location, and acquiring images and position data at the second location.

5. A method as claimed in claim 1, wherein the at least one camera module comprises a pair of cameras.

6. A method as claimed in claim 5, wherein the pair of cameras are separated by a distance greater than inter-ocular distance.

7. A method as claimed in claim 5, wherein each of the pair of cameras simultaneously capture an image of the projected array in the scene from the first and second positions.

8. A method as claimed in claim 1, wherein the two-dimensional array of points comprises a grid or checkerboard.

11. A method as claimed in claim 1, being configured to be performed in one of pipelines, subsea structures, horizontal flowlines, vertical risers, and subsea production and processing equipment.

12. A method as claimed in claim 1, being configured to be performed for an internal well bore survey at high speed using full resolution single or dual sensors, laser and pulsed lighting.

9. A method of generating a 3D point cloud of an underwater scene, the point cloud comprising a set of three variable Cartesian coordinates relating to surface points in the scene, the coordinates being defined in relation to an origin, the method operating in a system comprising a time-of-flight laser ranging device, a beam adjustor, a camera module and a controller, the method comprising:

the time-of-flight laser ranging device measuring the range to a point in the scene by projecting a laser beam onto that point;

recording a range timestamp associated with that range measurement;

the camera module capturing an image of the laser beam projected onto the scene, recording an image id associated with the captured image;

analysing the image to identify a horizontal and vertical coordinate of the laser beam in the image,

combining the range measurement with horizontal and vertical coordinates having an image time stamp that matches the range time stamp of the range measurement so as to form a three variable Cartesian coordinate; and

adjusting the location of the projected laser beam; and

repeating the steps above.

10. A method as claimed in claim 9, wherein the steps of measuring the range, recording a range time stamp, capturing an image and recording an image time stamp are repeated approximately 100 times per second.