IP Library › Granted Patent US 11,138,750
Granted Patent B2
US 11,138,750 · App. 16/591,016 · Granted Oct 5, 2021

Distance estimating method and apparatus

Inventors: Jae-woo Lee (Hwaseong-si, KR); Ja-hoo Koo (Seoul, KR); Won-ju Lee (Suwon-si, KR); Yoon-suk Hyun (Suwon-si, KR)
Assignee: SAMSUNG ELECTRONICS CO., LTD.
G06T7/536G06T7/20G06T7/60G06T7/70G06T2207/30256
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 11,138,750
App. No.
16/591,016
Filed
Oct 2, 2019
Granted
Oct 5, 2021
Kind
B2
Art Unit
2488
USPC
348/135
Abstract

A distance estimating method and apparatus are provided. The method includes acquiring a real target length corresponding to target points of an object, calculating, from an input image, an image target length corresponding to the real target length, and estimating a distance from a distance estimating apparatus to the object, based on the real target length, the image target length, and a focal length thereof.

Claims (59)

1. A distance estimating method comprising:

acquiring a real target length corresponding to target points of an object;

calculating, from an input image, an image target length corresponding to the real target length; and

estimating a distance from a distance estimating apparatus to the object, based on the real target length, the image target length, and a focal length thereof,

wherein the acquiring of the real target length includes calculating the real target length based on the image target length and a lane line appearing in the input image,

wherein the calculating the image target length includes:

detecting, from the input image, a bounding box corresponding to the object; and

determining the image target length from the bounding box,

wherein the calculating the real target length includes:

calculating, from the input image, a ratio between the image target length and an image lane width corresponding to a lane where the object is located; and

estimating the real target length from a real lane width based on the ratio,

wherein the calculating the ratio includes:

estimating the lane line for the lane where the object is located in the input image; and

determining the image lane width based on an intersection point between the lane line that is estimated and an extension line that extends from a base line of the bounding box toward the lane line,

wherein the estimating the real target length includes estimating the real lane width based on a projection image obtained by projecting the input image on a ground surface, and

wherein the estimating the real target length includes updating, in response to a case where a first point at which the lane line corresponding to the lane where the object is located is identified in a current frame is closer than a second point identified in a previous frame, the real lane width based on the first point.

2. The distance estimating method of claim 1 , wherein the estimating the lane line includes detecting the lane line by interpolating portions corresponding to a same line in the input image.

3. The distance estimating method of claim 1 , wherein the calculating the image target length includes:

estimating a pose of the object from the input image; and

determining the image target length based on the pose, and

wherein the image lane width is determined based on the pose of the object in the input image.

4. The distance estimating method of claim 1 , wherein the bounding box is a three-dimensional (3D) bounding box corresponding to the object; and

the image target length is determined based on the 3D bounding box, and

wherein the base line is of a rear face constituting the 3D bounding box.

5. The distance estimating method of claim 1 , wherein the estimating the real lane width includes:

identifying lane lines corresponding to the lane where the object is located in the projection image; and

estimating the real lane width based on a horizontal interval between boundary lines within a threshold distance from the distance estimating apparatus among the lane lines that are identified.

6. The distance estimating method of claim 1 , wherein the estimating the real target length includes acquiring, from map data corresponding to a current road, the real lane width of the lane where the object is located.

7. The distance estimating method of claim 1 , wherein the calculating the ratio includes calculating a pixel ratio between a first pixel distance corresponding to an interval between lane lines in the input image and a second pixel distance corresponding to a width of the object.

8. The distance estimating method of claim 1 , wherein the calculating the ratio includes excluding the real target length of the object that is estimated in response to a case where a width change of the lane where the object is located exceeds a threshold change.

9. The distance estimating method of claim 1 , wherein the acquiring of the real target length includes:

tracking the object allocated a previous target length acquired in the previous frame; and

determining a current target length of the object that is tracked in the current frame as the previous target length acquired in the previous frame.

10. The distance estimating method of claim 9 , wherein the tracking the object includes:

estimating the previous target length of the object based on an additional sensor in the previous frame; and

allocating the previous target length that is estimated to the object.

11. The distance estimating method of claim 1 , wherein the acquiring the real target length includes:

determining object dimension information corresponding to the object based on a visual appearance of the object represented in the input image; and

acquiring the real target length from the object dimension information.

12. The distance estimating method of claim 1 , wherein the estimating the distance includes calculating the distance from the focal length based on a ratio between the real target length and the image target length.

13. A non-transitory computer-readable recording medium having recorded thereon at least one computer program comprising at least one instruction for performing the distance estimating method of claim 1 .

14. A distance estimating apparatus comprising:

an image sensor configured to acquire an input image; and

a processor configured to:

acquire a real target length corresponding to target points of an object;

calculate, from the input image, an image target length corresponding to the real target length; and

estimate a distance from the distance estimating apparatus to the object, based on the real target length, the image target length, and a focal length thereof,

wherein the acquiring of the real target length includes calculating the real target length based on the image target length and a lane line appearing in the input image,

wherein the calculating the image target length includes:

detecting, from the input image, a bounding box corresponding to the object; and

determining the image target length from the bounding box,

wherein the calculating the real target length includes:

calculating, from the input image, a ratio between the image target length and an image lane width corresponding to a lane where the object is located; and

estimating the real target length from a real lane width based on the ratio,

wherein the calculating the ratio includes:

estimating the lane line for the lane where the object is located in the input image; and

determining the image lane width based on an intersection point between the lane line that is estimated and an extension line that extends from a base line of the bounding box toward the lane line,

wherein the estimating the real target length includes estimating the real lane width based on a projection image obtained by projecting the input image on a ground surface, and

wherein the estimating the real target length includes updating, in response to a case where a first point at which the lane line corresponding to the lane where the object is located is identified in a current frame is closer than a second point identified in a previous frame, the real lane width based on the first point.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 2, 2019
From: LEE, JAE-WOO; KOO, JA-HOO; LEE, WON-JU; HYUN, YOON-SUK
To: SAMSUNG ELECTRONICS CO., LTD.
Reel/Frame 050606/0312 →
Priority Claims (1)
KR 10-2018-0120158 · Oct 10, 2018 · national
Continuity (1)
Related Publication 20200118283A1 · Apr 16, 2020
Cited By (2)
US 12,601,586 US 12,694,547