IP Library Granted Patent US 11,087,495
Granted Patent B2
US 11,087,495 · App. 16/601,398 · Granted Aug 10, 2021

Method for calibration

Inventors: Jaemin Seo (Gyeonggi-do, KR); Seokmin Kim (Gyeonggi-do, KR); Seungkeol Kim (Gyeonggi-do, KR)
Assignee: TmaxData Co., Ltd.
G06T7/80G06K9/00523G06K9/3208G06T7/70G06K2209/23G06T2207/30244G06T2207/30252
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,087,495
App. No.
16/601,398
Granted
Aug 10, 2021
Kind
B2
Abstract

According to an exemplary embodiment of the present disclosure, a computer program stored in a computer readable storage medium is disclosed. The computer program causes one or more processors to perform operations below for performing calibration when the one or more processors are executed, and the operations may include: an operation of inputting an image to an object detection model and performing object detection by using the object detection model; an operation of acquiring bounding box information on the detected object; and an operation of performing calibration on an image acquisition device based on at least a part of the bounding box information.

Claims (59)

1. A non-transitory computer readable medium including a computer program, the computer program causing a computer to perform operations for calibration when executed by one or more processors, the operations including:

inputting an image into an object detection model and performing object detection using the object detection model;

obtaining bounding box information for the detected object; and

performing calibration on an image acquisition device based at least in part on the bounding box information,

wherein the bounding box information includes 3D bounding box information including a rotation angle information of a coordinate system of the image acquisition device,

wherein the rotation angle information includes at least one of azimuth of the coordinate system of the image acquisition device or elevation of the coordinate system of the image acquisition device.

2. The non-transitory computer readable medium according to claim 1 , wherein the object includes a vehicle object.

3. The non-transitory computer readable medium according to claim 1 , wherein the 3D bounding box information includes information of a coordinate system of the object set based on detailed component information of the object.

4. The non-transitory computer readable medium according to claim 1 , wherein the object detection model includes:

a convolutional layer to extract a feature of the image;

a first fully connected layer for obtaining an azimuth range of a coordinate system of the image acquisition device by receiving the extracted feature; and

a second fully connected layer for obtaining an azimuth of a coordinate system of the image acquisition device within the azimuth range.

5. The non-transitory computer readable medium according to claim 1 , wherein the object detection model includes:

a convolutional layer to extract a feature of the image;

a third fully connected layer for obtaining an elevation angle range of a coordinate system of the image acquisition device by inputting the extracted feature; and

a fourth fully connected layer for obtaining an elevation angle of a coordinate system of the image acquisition device within the elevation angle range.

6. The non-transitory computer readable medium according to claim 1 , wherein the performing calibration on an image acquisition device based at least in part on the bounding box information includes:

computing a distance between an image acquisition device and a reference point based on a respective bounding box information for at least one object;

selecting a reference object that is a reference for performing calibration based on predetermined criteria; and

performing calibration based on the reference object.

7. The non-transitory computer readable medium according to claim 6 , wherein the computing a distance between an image acquisition device and a reference point based on a respective bounding box information for at least one object includes,

computing a scale that is the ratio of the size of an object in a coordinate system of an image and the size of an object in a coordinate system of an image acquisition device.

8. The non-transitory computer readable medium according to claim 6 , wherein the predetermined criteria include at least one of:

a first criterion determined based on a statistic of a distance between the image acquisition device and a reference point computed based on respective bounding box information for at least one object; or

a second criterion determined based on the difference between an actual distance between an image acquisition device and a reference point and a computed distance between an image acquisition device and a reference point.

9. The non-transitory computer readable medium according to claim 6 , wherein the computing a distance between an image acquisition device and a reference point based on a respective bounding box information for at least one object includes:

converting a coordinate system of an image to a coordinate system of an image acquisition device; and

computing a distance between an image acquisition device and a reference point based on the coordinate system of an image acquisition device.

10. The non-transitory computer readable medium according to claim 9 , wherein the converting a coordinate system of an image to a coordinate system of an image acquisition device includes converting a coordinate system of an image to a coordinate system of an image acquisition device by using a first matrix including image acquisition device information and a second matrix including rotation angle information.

11. The non-transitory computer readable medium according to claim 10 , wherein the converting a coordinate system of an image to a coordinate system of an image acquisition device by using a first matrix including image acquisition device information and a second matrix including rotation angle information includes:

computing an intermediate result value by calculating a coordinate system of an image and an inverse of the first matrix; and

converting a coordinate system of an image to a coordinate system of an image acquisition device by calculating the computed intermediate result value and the inverse of the second matrix.

12. The non-transitory computer readable medium according to claim 1 , wherein the performing calibration on an image acquisition device based at least in part on the bounding box information includes:

computing a center coordinate of an object based on a respective bounding box information for at least one object;

determining a reference plane which is a reference for performing calibration based on the center coordinate of at least one object; and

performing calibration based on the reference plane.

13. The non-transitory computer readable medium according to claim 12 , wherein the center coordinate includes at least one of:

a first coordinate component computed based on the coordinate and scale of an object in a coordinate system of an image; or

a second coordinate component that includes a distance between the image acquisition device and a plane that includes the center coordinate of an object.

14. The non-transitory computer readable medium according to claim 13 ,

wherein the determining a reference plane which is a reference for performing calibration based on the center coordinate of the at least one object includes:

computing a third matrix comprising the first coordinate component at the center coordinates of the at least one object

computing a fourth matrix including the second coordinate component at the center coordinates of the at least one object; and

determining the reference plane based on the third matrix and the fourth matrix.

15. A method for calibration, the method comprising:

inputting an image into an object detection model and performing object detection using the object detection model;

obtaining bounding box information for the detected object; and

performing calibration on an image acquisition device based at least in part on the bounding box information,

wherein the bounding box information includes 3D bounding box information including a rotation angle information of a coordinate system of the image acquisition device,

wherein the rotation angle information includes at least one of azimuth of the coordinate system of the image acquisition device or elevation of the coordinate system of the image acquisition device.

16. A computing device comprising:

a processor including one or more cores; and

a memory;

wherein the processor is configured to:

input an image into an object detection model and perform object detection using the object detection model;

obtain bounding box information for the detected object; and

perform calibration on an image acquisition device based at least in part on the bounding box information,

wherein the bounding box information includes 3D bounding box information including a rotation angle information of a coordinate system of the image acquisition device,

wherein the rotation angle information includes at least one of azimuth of the coordinate system of the image acquisition device or elevation of the coordinate system of the image acquisition device.

Assignments (2)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 12, 2022
From: TMAXDATACO., LTD. (ALSO KNOWN AS TMAXDATA CO., LTD. AND TMAX DATA CO., LTD.)
To: TMAXTIBERO CO., LTD.
Reel/Frame 060635/0387 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 14, 2019
From: SEO, JAEMIN; KIM, SEOKMIN; KIM, SEUNGKEOL
To: TMAXDATA CO., LTD.
Reel/Frame 050709/0039 →
Priority Claims (1)
KR 10-2019-0076563 · Jun 26, 2019 · national
Continuity (1)
Related Publication 20200410716A1 · Dec 31, 2020