IP Library Granted Patent US 12670597
Granted Patent B2
US 12670597 · App. 17/788,682 · Granted Jun 30, 2026

Systems and methods for efficiently sensing collision threats

Inventors: Cedric Cocaud (Sunnyvale, CA); Arne Stoschek (Palo Alto, CA); Harvest Zhang (Sunnyvale, CA); Alexander Dean Naiman (Santa Clara, CA)
Assignee: A{circumflex over ( )}3 BY AIRBUS, LLC
G06T7/11G01S13/935G06T7/136G06V10/25G06V10/507G06V20/13G06T2207/10024G06T2207/10028G06T2207/10032G06T2207/20081G06T2207/20084G06T2207/30261
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 12670597
App. No.
17/788,682
Filed
Jun 23, 2022
Granted
Jun 30, 2026
Kind
B2
Art Unit
2676
USPC
382/103
Abstract

A system for efficiently sensing collision threats has an image sensor configured to capture an image of a scene external to a vehicle. The system is configured to then identify an area of the image that is associated with homogeneous sensor values and is thus likely devoid of collision threats. In order to reduce the computational processing required for detecting collision threats, the system culls the identified area from the image, thereby conserving the processing resources of the system.

Claims (46)

1 . A vehicular system for sensing collision threats, comprising:

a vehicle;

a first image sensor coupled to the vehicle and configured to capture a first image of a scene external to the vehicle;

a second sensor coupled to the vehicle and configured to provide sensor values based on reflections from the scene external to the vehicle;

at least one processor in communication with a non-transitory memory storing instructions, that when executed, cause the processor to:

identify a homogenous grouping of sensor values from the sensor values provided by the second sensor, wherein the homogenous grouping of sensor values correspond to a first geographic region;

identify an area of the first image associated with the first geographic region that is correlated to both (i) the homogenous grouping of sensor values provided by the second sensor and (ii) non-homogenous sensor values provided by the first image sensor;

responsive to identifying the area of the first image associated with both (i) the homogenous grouping of sensor values provided by the second sensor and (ii) the non-homogenous sensor values provided by the first sensor, cull the identified area from the first image thereby providing a filtered image;

process the filtered image to identify a collision threat within the filtered image; and

provide information indicative of the detected collision threat.

2 . The vehicular system of claim 1 , wherein the at least one processor implements a machine learning algorithm for processing the filtered image to identify the collision threat.

3 . The vehicular system of claim 1 , wherein the homogeneous grouping of sensor values correspond to pixels of the first image.

4 . The vehicular system of claim 1 , further comprising a second image sensor coupled to the vehicle and configured to capture a second image of the scene, wherein the homogeneous grouping of sensor values correspond to pixels of the second image.

5 . The vehicular system of claim 1 , wherein the second sensor is a radar sensor.

6 . The vehicular system of claim 1 , wherein the second sensor is a light detection and ranging (LiDAR) sensor.

7 . The vehicular system of claim 1 , wherein the at least one processor is configured to determine a value indicative of a homogeneity of the homogeneous grouping of sensor values and compare the value to a threshold.

8 . A vehicular system for sensing collision threats, comprising:

a vehicle;

a first image sensor coupled to the vehicle and configured to capture a first image of a scene external to the vehicle;

a second sensor coupled to the vehicle and configured to provide sensor values based on reflections from the scene external to the vehicle;

at least one processor configured to receive the first image, the at least one processor programmed with instructions that, when executed by the at least one processor, cause the at least one processor to:

identify a homogenous grouping of sensor values from the sensor values provided by the second sensor, wherein the homogenous grouping of sensor values correspond to a first geographic region;

determine whether an area of the first image is associated with the first geographic region that is correlated to both (i) the homogenous grouping of sensor values provided by the second sensor and (ii) non-homogenous sensor values provided by the first image sensor;

in response to determining that the area of the first image is associated with both (i) the homogenous grouping of sensor values provided by the second sensor and (ii) the non-homogenous sensor values provided by the first sensor, cull the area from the first image, thereby providing a filtered image;

detect a collision threat for the vehicle based on the filtered image; and

provide information indicative of the detected collision threat.

9 . The vehicular system of claim 8 , wherein the homogeneous sensor values comprise pixels of the first image.

10 . The vehicular system of claim 8 , further comprising a second image sensor coupled to the vehicle and configured to capture a second image of the scene, wherein the homogeneous grouping of sensor values correspond to pixels of the second image.

11 . The vehicular system of claim 8 , wherein the second sensor is a radar sensor.

12 . The vehicular system of claim 8 , wherein the second sensor is a light detection and ranging (LiDAR) sensor.

13 . The vehicular system of claim 8 , wherein the at least one processor is configured to determine a value indicative of a homogeneity of the homogeneous grouping of sensor values and compare the value to a threshold.

14 . A method for sensing collision threats, comprising:

capturing a first image of a scene external to a vehicle with a first image sensor;

receiving a homogenous grouping of sensor values from a second sensor based on reflections from the scene external to the vehicle, wherein the homogenous grouping of sensor values correspond to a first geographic region;

identifying, with at least one processor, an area of the first image associated with the first geographic region that is correlated to both (i) the homogenous grouping of sensor values provided by the second sensor and (ii) non-homogenous sensor values provided by the first image sensor;

responsive to identifying the area of the first image associated with both (i) the homogenous grouping of sensor values provided by the second sensor and (ii) the non-homogenous sensor values provided by the first sensor, culling the area from the first image with the at least one processor;

analyzing the first image subsequent to the culling with the at least one processor;

detecting a collision threat to the vehicle with the at least one processor based on the analyzing; and

providing, with the at least one processor, information indicative of the detected collision threat.

15 . The method of claim 14 , wherein the homogeneous grouping of sensor values correspond to pixels of the first image.

16 . The method of claim 14 , further comprising capturing a second image of the scene with a second image sensor, wherein the homogeneous grouping of sensor values correspond to pixels of the second image.

17 . The method of claim 14 , wherein the second sensor is a radar sensor.

18 . The method of claim 14 , wherein the second sensor is a light detection and ranging (LiDAR) sensor.

19 . The method of claim 14 , further comprising:

determining a value indicative of a homogeneity of the homogeneous grouping of sensor values; and

comparing the value to a threshold.