Systems and methods for stereoscopic imaging

Patent metadata

Field	Value
Publication number	US-10896519-B2
Application number	US-201916240704-A
Country	US
Kind code	B2
Filing date	Jan 4, 2019
Priority date	Jul 6, 2016
Publication date	Jan 19, 2021
Grant date	Jan 19, 2021

Abstract

A method of block matching includes selecting a block configuration from a plurality of block configurations and matching a first point in a first image with a corresponding second point in a second image using the selected block configuration.

First claim

What is claimed is:

1. A method of block matching for providing depth information of an object, the method comprising:
setting a plurality of block configurations each corresponding to one block configuration calculator and each including one or more configuration parameters for defining blocks on images, the one or more configuration parameters including block sizes and block shapes;
selecting a first block configuration from the plurality of block configurations and selecting a second block configuration from the plurality of block configurations, a block size of the selected first block configuration being determined at least by a texture of the object on the blocks of the images, the texture of the object being identified by an image feature recognition, and the selected second block configuration having at least one configuration parameter different from the selected first block configuration;
choosing a first block of pixels around a first point of a first image based on the selected first block configuration and the selected second block configuration;
selecting a first candidate second point from a plurality of candidate second points of a second image and locating a second block of pixels around the first candidate second point based on the selected first block configuration and the selected second block configuration;
calculating, by a corresponding block configuration calculator of the selected first block configuration, a first matching cost for the selected first block configuration between the first point of the first image and the first candidate second point of the second image, and calculating, by a corresponding block configuration calculator of the selected second block configuration, a second matching cost for the selected second block configuration between the first point of the first image and the first candidate second point of the second image;
selecting a second candidate second point from the plurality of candidate second points of the second image, and calculating a third matching cost for the selected first block configuration and a fourth matching cost for the selected second block configuration, respectively, based on the second candidate second point; and
obtaining a cross-block comparison based on the first matching cost, the second matching cost, the third matching cost, and the fourth matching cost to match the first point in the first image with a corresponding second point in the second image, thereby improving an accuracy and reliability of the depth information to enhance a control of a mobile platform with respect to the object.

2. The method of claim 1 , wherein matching the first point with the corresponding second point comprises:
determining matching costs between the first point in the first image and the plurality of candidate second points in the second image; and
identifying the corresponding second point from the plurality of candidate second points based on the matching costs.

3. The method of claim 2 , where determining the matching costs between the first point and the plurality of candidate second points comprises determining matching costs between the first point in the first image and a plurality of candidate second points on a line of the second image.

4. The method of claim 2 , wherein determining the matching costs between the first point and the plurality of candidate second points comprises, for one candidate second point of the plurality of candidate second points:
determining individual matching costs between the first point and the candidate second point, each of the individual matching costs being determined based on one block configuration of the block configurations; and
determining the matching cost between the first point and the candidate second point based on the individual matching costs.

5. The method of claim 1 , further comprising:
determining an object distance based on a disparity between the first point of the first image and the corresponding second point of the second image.

6. The method of claim 1 , wherein matching the first point with the corresponding second point comprises determining individual matching costs between the first point and each of the plurality of candidate second points in the second image based on each of the block configurations to select multiple candidate second points respectively corresponding to the block configurations.

7. The method of claim 6 , further comprising:
obtaining candidate disparities between the first point and the selected multiple candidate second points, respectively; and
determining a disparity between the first point and the corresponding second point based on the candidate disparities.

8. A stereoscopic imaging system for providing depth information of an object, the system comprising:
a first imaging device configured to obtain a first image;
a second imaging device configured to obtain a second image; and
one or more processors configured to:
set a plurality of block configurations each corresponding to one block configuration calculator and each including one or more configuration parameters for defining blocks on images, the one or more configuration parameters including block sizes and block shapes;
select a first block configuration from the plurality of block configurations and select a second block configuration from the plurality of block configurations, a block size of the selected first block configuration being determined at least by a texture of the object on the blocks of the images, the texture of the object being identified by an image feature recognition, and the selected second block configuration having at least one configuration parameter different from the selected first block configuration;
choose a first block of pixels around a first point of the first image based on the selected first block configuration and the selected second block configuration;
select a first candidate second point from a plurality of candidate second points of the second image and locate a second block of pixels around the first candidate second point based on the selected first block configuration and the selected second block configuration;
calculate, by a corresponding block configuration calculator of the selected first block configuration, a first matching cost for the selected first block configuration between the first point of the first image and the first candidate second point of the second image, and calculate, by a corresponding block configuration calculator of the selected second block configuration, a second matching cost for the selected second block configuration between the first point of the first image and the first candidate second point of the second image;
select a second candidate second point from the plurality of candidate second points of the second image, and calculate a third matching cost for the selected first block configuration and a fourth matching cost for the selected second block configuration, respectively, based on the second candidate second point; and
obtain a cross-block comparison based on the first matching cost, the second matching cost, the third matching cost, and the fourth matching cost to match the first point in the first image with a corresponding second point in the second image, thereby improving an accuracy and reliability of the depth information to enhance a control of a mobile platform with respect to the object.

9. The stereoscopic imaging system of claim 8 , wherein the one or more processors are further configured to:
determine matching costs between the first point in the first image and the plurality of candidate second points in the second image; and
identify the corresponding second point from the plurality of candidate second p

Assignees

• Sz Dji Technology Co Ltd

Inventors

• Yang Kang
• Gao Mingming

Classifications

• H04N13/128

  Adjusting depth or disparity · CPC title

• G06T7/593 Primary

  from stereo images · CPC title

• H04N2013/0081

  Depth or disparity estimation from stereoscopic image signals · CPC title

• H04N13/239

  using two two-dimensional [2D] image sensors having a relative position equal to or related to the interocular distance (H04N13/243 takes precedence) · CPC title

• G06T2207/20021

  Dividing image into blocks, subimages or windows · CPC title