Data processing method, apparatus, and system

Patent metadata

Field	Value
Publication number	US-10909690-B2
Application number	US-201816211978-A
Country	US
Kind code	B2
Filing date	Dec 6, 2018
Priority date	Jun 7, 2016
Publication date	Feb 2, 2021
Grant date	Feb 2, 2021

Abstract

A data processing method includes reading first to-be-processed data from a first storage unit into a second storage unit, computing a processing result according to the first to-be-processed data stored in the second storage unit, and, while computing the processing result, reading second to-be-processed data from the first storage unit. The first to-be-processed data is at a first layer of one or more image pyramids having a plurality of layers, and the second to-be-processed data is at a second layer of the one or more image pyramids.

First claim

What is claimed is:

1. A data processing method, comprising:
reading first to-be-processed data from a first storage unit into a second storage unit, wherein the first to-be-processed data is at a first layer of one or more image pyramids having a plurality of layers;
computing a processing result according to the first to-be-processed data stored in the second storage unit; and
while computing the processing result, reading second to-be-processed data from the first storage unit, wherein the second to-be-processed data is at a second layer of the one or more image pyramids;
wherein:
the one or more image pyramids includes a first image pyramid including a first original pyramid of a first image and a second image pyramid including a second original pyramid of a second image;
the second to-be-processed data includes:
data within a first tracking window at the second layer of the first image pyramid, wherein a position of the first tracking window is determined according to a position of a tracking point at the second layer of the first original pyramid; and
data within a second tracking window at the second layer of the second image pyramid, wherein a position of the second tracking window is determined according to the processing result; and

the processing result comprises an offset correction of the tracking point at the first layer and a predicted offset at the second layer.

2. The method according to claim 1 , wherein computing the processing result comprises:
before completing reading the first to-be-processed data, starting to compute the processing result through a plurality of iteration processes according to at least a portion of the first to-be-processed data stored in the second storage unit.

3. The method according to claim 2 , wherein:
the first to-be-processed data comprises data at the first layer that is used in one or more of the plurality of iteration processes; and
reading the first to-be-processed data comprises reading the first to-be-processed data with a single reading operation.

4. The method according to claim 3 , wherein:
reading the second to-be-processed data from the first storage unit comprises:
reading data at the second layer of the first image pyramid from the first storage unit; and
after computing the processing result, reading data at the second layer of the second image pyramid from the first storage unit according to the processing result.

5. The method according to claim 4 , wherein computing the processing result further comprises computing the processing result according to gradient data of the first image at the first layer.

6. The method according to claim 5 , wherein:
the first image further comprises a gradient pyramid; and
reading the first to-be-processed data further comprises reading the gradient data into the second storage unit, the gradient data comprising data at the first layer of the gradient pyramid within the first tracking window.

7. The method according to claim 5 , further comprising:
calculating the gradient data of the first image according to data of the first original pyramid within the first tracking window at the first layer stored in the second storage unit.

8. The method according to claim 4 , wherein the data within the second tracking window at the second layer of the second image pyramid data comprises data of the second original pyramid at the second layer used in one or more of the plurality of iteration processes.

9. The method according to claim 1 , further comprising:
obtaining a positional offset of the tracking point from the first image to the second image according to the offset correction and the predicted offset of the tracking point at a lowest layer.

10. A data processing apparatus, comprising:
a first storage unit and a second storage unit;
a memory storing one or more computer programs; and
a processor coupled to the memory and configured to execute the one or more computer programs to:
read first to-be-processed data from the first storage unit into the second storage unit, wherein the first to-be-processed data is at a first layer of one or more image pyramids having a plurality of layers;
compute a processing result according to the first to-be-processed data stored in the second storage unit; and
while computing the processing result, read second to-be-processed data from the first storage unit, wherein the second to-be-processed data is at a second layer of the one or more image pyramids;

wherein:
the one or more image pyramids includes a first image pyramid including a first original pyramid of a first image and a second image pyramid including a second original pyramid of a second image;
the second to-be-processed data includes:
data within a first tracking window at the second layer of the first image pyramid, wherein a position of the first tracking window is determined according to a position of a tracking point at the second layer of the first original pyramid; and
data within a second tracking window at the second layer of the second image pyramid, wherein a position of the second tracking window is determined according to the processing result; and
the processing result comprises an offset correction of the tracking point at the first layer and a predicted offset at the second layer.

11. The data processing apparatus according to claim 10 , wherein:
the processor is further configured to execute the one or more computer programs to, before completing reading the first to-be-processed data, start computing the processing result through a plurality of iteration processes according to at least a portion of the first to-be-processed data stored in the second storage unit.

12. The data processing apparatus according to claim 11 , wherein:
the first to-be-processed data comprises data at the first layer that is used in one or more of the plurality of iteration processes; and
the processor is further configured to execute the one or more computer programs to read the first to-be-processed data with a single reading operation.

13. The data processing apparatus according to claim 12 , wherein:
the processor is further configured to execute the one or more computer programs to:
read data at the second layer of the first image pyramid from the first storage unit; and
after computing the processing result, read data at the second layer of the second image pyramid from the first storage unit according to the first processing result.

14. The data processing apparatus according to claim 13 , wherein the processor is further configured to execute the one or more computer programs to compute the processing result according to gradient data of the first image at the first layer.

15. A non-transitory computer-readable storage medium storing computer program instructions executable by at least one processor to:
read first to-be-processed data from a first storage unit into a second storage unit, wherein the first to-be-processed data is at a first layer of one or more image pyramids having a plurality of layers;
compute a processing result according to the first to-be-processed data stored in the second storage unit; and
while computing the processing result, read second to-be-processed data from the first storage unit, wherein the second to-be-processed data is at a second layer of the one or more image pyramids;
wherein:
the one or more image pyramids includes a first image pyramid including a first original pyramid of a first image and a second image pyramid including a second original pyramid of a second image;
the second to-be-processed data includes:
data within a

Assignees

• Sz Dji Technology Co Ltd

Inventors

• Gao Mingming
• Yang Kang

Classifications

• G06T2207/20016

  Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform · CPC title

• G06T7/207 Primary

  for motion estimation over a hierarchy of resolutions (multi-resolution motion estimation or hierarchical motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/53) · CPC title

• G06T7/269

  using gradient-based methods · CPC title

• G06T7/246 Primary

  using feature-based methods, e.g. the tracking of corners or segments · CPC title