IP Library › Granted Patent US 12,738,074
Granted Patent B1
US 12,738,074 · App. 19/426,954 · Granted Sep 15, 2026

Systems and methods for traffic sign detection

Inventors: Jorge González Núñez (Madrid, ES); Álvaro Alcalde Cid (Madrid, ES); Daniel Pérez Rubio (Azuqueca de Henares, ES); Mohammed Sohail Siddique (Milton, CA)
Assignee: Geotab Inc.
G06V20/582G01C21/3815G01C21/3841G06V10/70G06V20/588G06V20/63G06V30/10
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 12,738,074
App. No.
19/426,954
Granted
Sep 15, 2026
Kind
B1
Abstract

A method and system are provided for enhancing map data of a road network by automatically generating a dataset of traffic signs. The method includes receiving telematics data from a plurality of vehicles and receiving map data comprising a plurality of road edges. A point of interest associated with at least one road edge is determined based on a vehicle movement metric derived from the telematics data, such as speed- or stop-related patterns. A sequence of one or more images associated with the point of interest is obtained from an image capture device mounted to a vehicle traversing the point of interest. At least one image from the sequence is analyzed using an image-processing machine learning model to detect a traffic sign, and the detected traffic sign is used to enhance the map data. The approach focuses image analysis on selected portions of the road network and reduces data processing by limiting analysis to relevant images, thereby enabling efficient generation of a reliable, updated dataset of traffic signs.

Claims (32)

1 . A method for enhancing map data of a road network, the method comprises operating at least one processor to:

receive telematics data originating from a plurality of telematics devices installed in a plurality of vehicles;

receive the map data wherein the map data comprising a plurality of road edges,

determine a point of interest associated with at least one road edge of the plurality of road edges, wherein the determination of the point of interest is based on a vehicle movement metric derived from the telematics data of the plurality of vehicles traversing the at least one road edge;

determine a sequence of one or more images associated with the determined point of interest wherein the sequence of one or more images is captured by an image capture device positioned at a vehicle traversing the determined point of interest; and

analyze at least one image from the sequence of one or more images using a machine learning model to detect a traffic sign therein for using the detected traffic sign to enhance the map data.

2 . The method according to claim 1 wherein the determination of the point of interest comprises a determination of a road intersection with the associated at least one road edge.

3 . The method according to claim 1 wherein the determination of the point of interest comprises a determination of a predefined geographical area with the associated at least one road edge, wherein the predefined geographical area is defined in relation to an infrastructural area.

4 . The method according to claim 1 wherein the vehicle movement metric is indicative of a pattern of vehicle movements on the at least one road edge.

5 . The method according to claim 1 wherein the vehicle movement metric is indicative of a variability of vehicle speeds on the at least one road edge and calculated as a speed entropy.

6 . The method according to claim 5 wherein the speed entropy is calculated for each of the at least one road edges and a road edge is assigned to a point of interest when the speed entropy indicates a high variability.

7 . The method according to claim 5 wherein the speed entropy is calculated for each of a plurality of vehicle types and for a particular traffic time period.

8 . The method according to claim 1 wherein the vehicle movement metric is indicative of a frequency of vehicle braking events or vehicle stops derived from the telematics data for the at least one road edge.

9 . The method according to claim 1 wherein the sequence of one or more images associated with the determined point of interest is determined by identifying a vehicle traversing the determined point of interest.

10 . The method according to claim 9 wherein the sequence of one or more images associated with the determined point of interest is determined by identifying a time point the vehicle enters the determined point of interest and a time point the vehicle leaves the determined point of interest.

11 . The method according to claim 1 wherein the machine learning model is an image-processing machine learning model.

12 . The method according to claim 11 wherein the machine learning model is configured to process the traffic sign detection and a text related to the traffic sign as a one-stage process.

13 . The method according to claim 11 wherein the machine learning model is configured to process the traffic sign detection and a text related to the traffic sign as a two-stage process comprising detecting the traffic sign in a first stage and recognizing a text related to the traffic sign in a second stage.

14 . The method according to claim 1 wherein the step of analyzing the at least one image from the sequence of one or more images comprises extracting at least one key frame from the sequence of one or more images and applying the machine learning model to detect the traffic sign in the at least one key frame.

15 . A system for enhancing map data of a road network, the system comprising:

at least one data storage operable to store telematics data originating from a plurality of telematics devices installed in a plurality of vehicles; and

at least one processor in communication with the at least one data storage, the at least one processor operable to:

receive telematics data originating from a plurality of telematics devices installed in a plurality of vehicles;

receive the map data wherein the map data comprising a plurality of road edges,

determine a point of interest associated with at least one road edge of the plurality of road edges, wherein the determination of the point of interest is based on a vehicle movement metric derived from the telematics data of the plurality of vehicles traversing the at least one road edge;

determine a sequence of one or more images associated with the determined point of interest wherein the sequence of one or more images is captured by an image capture device positioned at a vehicle traversing the determined point of interest; and

analyze at least one image from the sequence of one or more images using a machine learning model to detect a traffic sign therein for using the detected traffic sign to enhance the map data.

16 . The system according to claim 15 , wherein the at least one processor is further operable to determine the point of interest by determining a road intersection with the associated at least one road edge.

17 . The system according to claim 15 wherein the wherein the at least one processor is further operable to determine a predefined geographical area with the associated at least one road edge, wherein the predefined geographical area is defined in relation to an infrastructural area.

18 . The system according to claim 15 wherein the vehicle movement metric is indicative of a variability of vehicle speeds on the at least one road edge and calculated as a speed entropy.

19 . The system according to claim 18 wherein the speed entropy is calculated for each of the at least one road edges and a road edge is assigned to a point of interest when the speed entropy indicates a high variability.

20 . The system according to claim 15 wherein the at least one processor is further operable to analyze the at least one image from the sequence of one or more images by extracting at least one key frame from the sequence of one or more images and applying the machine learning model to detect the traffic sign in the at least one key frame.

Assignments (3)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 19, 2025
From: SIDDIQUE, MOHAMMED SOHAIL
To: GEOTAB INC.
Reel/Frame 073275/0428 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 19, 2025
From: NÚÑEZ, JORGE GONZÁLEZ; ALCALDE CID, ÁLVARO; PÉREZ RUBIO, DANIEL
To: GEOTAB GMBH SUCURSAL EN ESPANA
Reel/Frame 073275/0537 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 19, 2025
From: GEOTAB GMBH SUCURSAL EN ESPANA
To: GEOTAB INC.
Reel/Frame 073275/0566 →
Continuity (1)
Provisional Application 63845556 · Jul 17, 2025
References Cited (14)
US 9195895B1 · Kapach · 2015 [cited by examiner]
US 10042359B1 · Konrardy · 2018 [cited by examiner]
US 10402665B2 · Kapach · 2019 [cited by examiner]
US 10452069B2 · Stein · 2019 [cited by examiner]
US 10599155B1 · Konrardy · 2020 [cited by examiner]
US 11113545B2 · Kundu · 2021 [cited by examiner]
US 11385058B2 · Prabhakar · 2022 [cited by examiner]
US 11860979B2 · Cardona · 2024 [cited by examiner]
US 12169985B1 · Gataric · 2024 [cited by examiner]
US 12306010B1 · Rommel · 2025 [cited by examiner]
US 12511994B2 · Karakayis · 2025 [cited by examiner]
US 12525029B2 · Singh · 2026 [cited by examiner]
Qi et al. (CN 117372633 ARoad Boundary Generating Method and Device Based on Three-dimensional Road Surface); date published Jan. 9, 2024. [cited by examiner]
Gao et al. (CN 117824695 A, Multi-dimensional Intelligent Driving Path Planning System); date published Apr. 5, 2024. [cited by examiner]