Systems and methods for calibrating vehicle following distance determination

Patent metadata

Field	Value
Publication number	US-12596196-B2
Application number	US-202519307170-A
Country	US
Kind code	B2
Filing date	Aug 22, 2025
Priority date	Sep 6, 2024
Publication date	Apr 7, 2026
Grant date	Apr 7, 2026

Abstract

Systems and methods for validating following distance detection models are discussed. A LIDAR unit and an image capture device are installed at a vehicle. LIDAR data from the LIDAR unit is used to identify a lead vehicle and for determination of a true following distance to the lead vehicle. An image-based following distance detection model is applied on image data captured by the image capture device. Following distance as determined by the image-based following distance model is compared to the true distance from the LIDAR data, to assess validity of the image-based following distance model.

First claim

What is claimed is:

1 . A system comprising:
at least one processor; at least one non-transitory processor-readable storage medium storing processor executable instructions which, when executed by the at least one processor, cause the system to: access an image-based distance value corresponding to an image frame, the image-based distance value indicating a following distance between a first vehicle and a second vehicle, as output by an image-based distance determination model which analyzes the image frame as captured by an image capture device positioned at the second vehicle; access LIDAR data captured by a LIDAR sensor positioned at the second vehicle; correlate, by the at least one processor, data for a LIDAR frame of the LIDAR data to a timestamp of the image frame; apply, by the at least one processor, a clustering algorithm to the LIDAR frame to identify a cluster of LIDAR data points representing the first vehicle; determine, by the at least one processor, a LIDAR-based distance between the second vehicle and the cluster of LIDAR data points; determine, by the at least one processor, a difference between the LIDAR-based distance and the image-based distance value; if the difference between the LIDAR-based distance and the image-based distance value is within a threshold ratio of the LIDAR-based distance, label the image-based distance value as true; and if the difference between the LIDAR-based distance and the image-based distance value exceeds the threshold ratio of the LIDAR-based distance, label the image-based distance value as false.

2 . The system of claim 1 , wherein the processor-executable instructions which cause the system to correlate the LIDAR frame to the timestamp of the image frame cause the at least one processor to:
synchronize the LIDAR data to a server clock to which the image frame is synchronized; and selecting the LIDAR frame from a plurality LIDAR frames data, based on the LIDAR frame having a timestamp which is closest to a timestamp of the image frame.

3 . The system of claim 1 , wherein the processor-executable instructions further cause the at least one processor to:
execute a ground surface identification model on the LIDAR frame to identify a ground surface plane; and calibrate the LIDAR frame to compensate for a non-zero angle between the LIDAR sensor and the ground surface plane.

4 . The system of claim 1 , wherein the processor-executable instructions further cause the at least one processor to:
identify a region of interest in the LIDAR frame positioned in front of the second vehicle.

5 . The system of claim 4 , wherein the processor executable instructions which cause the at least one processor to identify the region of interest cause the at least one processor to:
identify a lane of travel of the second vehicle based on the LIDAR frame; and identify the region of interest as being limited to the lane of travel.

6 . The system of claim 5 , wherein the processor-executable instructions which cause the at least one processor to identify the lane of travel of the second vehicle cause the at least one processor to identify lane markings represented in the LIDAR frame which delineate the lane of travel in which the second vehicle is positioned.

7 . The system of claim 1 , wherein the processor-executable instructions which cause the at least one processor to determine LIDAR-based distance between the second vehicle and the cluster of LIDAR data points cause the at least one processor to identify the LIDAR based distance as a distance between the second vehicle and a data point in the cluster of LIDAR data points which is closest to the second vehicle.

8 . The system of claim 1 , wherein the processor-executable instructions which cause the at least one processor to determine the LIDAR-based distance between the second vehicle and the cluster of LIDAR data points cause the at least one processor to determine the LIDAR-based distance as a distance between the second vehicle and an average or median position of the cluster of LIDAR data points.

9 . The system of claim 1 , wherein:
the processor-executable instructions which cause the system to access an image-based distance value corresponding to an image frame cause the system to: access, a plurality of image-based distance values, each image-based distance value indicating a respective following distance between a first vehicle and a second vehicle, as output by the image-based distance determination model which analyzes a respective image frame as captured by an image sensor positioned at the second vehicle; the processor-executable instructions which cause the at least one processor to correlate a LIDAR frame of the LIDAR data to a timestamp of the image frame cause the at least one processor to: correlate a respective frame of a plurality of frames of the LIDAR data to respective timestamps of the plurality of image frames; the processor-executable instructions which cause the at least one processor to apply a clustering algorithm to the LIDAR frame to identify a cluster of LIDAR data points representing the first vehicle cause the at least one processor to: apply the clustering algorithm to each LIDAR frame to identify a respective cluster of LIDAR data points representing the first vehicle for each LIDAR frame; the processor-executable instructions which cause the at least one processor to identify a LIDAR-based distance between the second vehicle and the cluster of LIDAR data points cause the at least one processor to: identify a respective LIDAR-based distance between the second vehicle and the respective cluster of LIDAR data points for each LIDAR frame; and the processor-executable instructions which cause the at least one processor to determine a difference between the LIDAR-based distance and the image-based distance value cause the at least one processor to: determine a respective difference between the LIDAR-based distance and the respective image-based distance value for each LIDAR frame.

10 . The system of claim 9 , wherein;
the processor-executable instructions which cause the at least one processor to, if the difference between the LIDAR-based distance and the image-based distance value is within a threshold ratio of the LIDAR-based distance, label the image-based distance value as true, cause the at least one processor to, for each LIDAR frame: if the respective difference between the respective LIDAR-based distance and the respective image-based distance value is within a threshold ratio of the respective LIDAR-based distance, label the respective image-based distance value as true; and the processor-executable instructions which cause the at least one processor to, if the difference between the LIDAR-based distance and the image-based distance value exceeds the threshold ratio of the LIDAR-based distance, label the image-based distance value as false, cause the at least one processor to, if the respective difference between the respective LIDAR-based distance and the respective image-based distance value exceeds the threshold ratio of the respective LIDAR-based distance, label the respective image-based distance value as false.

11 . The system of claim 10 , wherein the processor-executable instructions further cause the at least one processor to:
determine image-based distance value accuracy for the plurality of image frames as a percentage of image frames of the plurality of image frames where the respective image-based distance value is labelled as true.

12 . The system of claim 1 , further comprising the LIDAR sensor, wherein the processor-executable instructions which cause the system to access the LIDAR d

Assignees

• Geotab Inc

Inventors

• Saurav Shashank
• Siddique Javed
• Siddique Mohammed Sohail
• Tiwari Tuhin
• Qiao Donghao

Classifications

• G06T2207/30261

  Obstacle · CPC title

• G06T2207/30256

  Lane; Road marking · CPC title

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title

• G01S17/89

  for mapping or imaging · CPC title

• G06V2201/08

  Detecting or categorising vehicles · CPC title