# US 2022/0092314 A1 · Road Recognition Method and System Based on Seed Point

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0092314 A1
- Document ID: US20220092314A1
- Status: Published application
- Kind: A1
- Application number: 17/145,193
- Filed: 2021-01-08
- Published: 2022-03-24
- Earliest claimed: 2020-09-21
- Assignees: Nanjing University of Aeronautics and Astronautics
- Applicants: Nanjing University of Aeronautics and Astronautics
- Inventors: Jie Hao; Lei Zhang
- Web page: https://patentgrep.com/patent/US20220092314A1

## CPC

- G06F 16/583: using metadata automatically derived from the content
- G06V 20/56: exterior to a vehicle by using sensors mounted on the vehicle
- G06F 16/535: Filtering based on additional data, e.g. user or group profiles
- G06F 16/55: Clustering; Classification
- G06N 20/00: Machine learning
- G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI]
- G06V 10/30: Noise filtering
- G06V 20/588: {Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

The present disclosure relates to a road recognition method and system based on a seed point. The method includes: obtaining a remote sensing image; obtaining a grayscale image; inserting a seed point in the grayscale image; searching for four initial road boundary points using the seed point as a reference point; obtaining a smallest bounding rectangle search box formed by the four initial road boundary points; obtaining a plurality of candidate search boxes based on the obtained search boxes; determining whether a sum of squares of grayscale differences between the plurality of candidate search boxes and the obtained search box is greater than a preset threshold; and if yes, stopping searching and completing road recognition; or if not, selecting a new search box from the plurality of candidate search boxes, and retrieving a plurality of candidate search boxes based on the new search box.

## Claims

### Claim 1

A road recognition method based on a seed point, comprising:
obtaining a remote sensing image;
preprocessing the remote sensing image to obtain a grayscale image;
inserting a seed point in the grayscale image, wherein the seed point is a monitoring location point in a road to be identified;
searching for four initial road boundary points using the seed point as a reference point;
obtaining a smallest bounding rectangle search box formed by the four initial road boundary points;
rotating the obtained search box at different angles, and moving rotated search boxes to obtain a plurality of candidate search boxes;
determining whether a sum of squares of grayscale differences between the plurality of candidate search boxes and the obtained search box is greater than a preset threshold; and
if yes, stopping searching and completing road recognition; or
if not, selecting a new search box from the plurality of candidate search boxes, and retrieving a plurality of candidate search boxes based on the new search box.

### Claim 2 (depends on 1)

The road recognition method according to claim 1, wherein the searching for four initial road boundary points using the seed point as a reference point is specifically:
separately searching for pixels in α° direction and (α+180)° direction of the seed point;
determining whether a gray value difference between a current searched pixel and a previous pixel is greater than a preset threshold; and
if yes, using the current pixel as one initial road boundary point, and then changing the search direction with the seed point as the reference point to continue searching for a pixel; or
if not, directly changing the search direction with the seed point as the reference point to continue searching for a pixel.

### Claim 3 (depends on 2)

The road recognition method according to claim 2, wherein search directions obtained after changing are (α+45)° and (α+225)°, (α+90)° and (α+270)°, and (α+135)° and (α+315)°.

### Claim 4 (depends on 1)

The road recognition method according to claim 1, wherein the rotating the obtained search box at different angles, and moving rotated search boxes to obtain a plurality of candidate search boxes is specifically:
rotating the obtained search box with a center point of the obtained search box as a center, and advancing the rotated search box by a side length of the obtained search box to obtain a plurality of candidate search boxes advanced at different rotation angles.

### Claim 5 (depends on 1)

The road recognition method according to claim 1, wherein the selecting a new search box from the plurality of candidate search boxes is specifically:
selecting a candidate search box from the plurality of candidate search boxes as a new search box, wherein a sum of squares of grayscale differences between the candidate search box and the obtained search box is smallest.

### Claim 6 (depends on 1)

The road recognition method according to claim 1, wherein a search step value for searching the four initial road boundary points is one pixel, and a search range is 10 pixels.

### Claim 7 (depends on 1)

The road recognition method of claim 1, wherein the preprocessing the remote sensing image to obtain a grayscale image is specifically:
performing geometric correction, radiometric calibration, and graying on the remote sensing image to obtain the grayscale image.

### Claim 8

A road recognition system based on a seed point, comprising:
an image obtaining module, configured to obtain a remote sensing image;
an image preprocessing module, configured to preprocess the remote sensing image to obtain a grayscale image;
a seed point insertion module, configured to insert a seed point in the grayscale image, wherein the seed point is a monitoring location point in a road to be identified;
a search module, configured to search for four initial road boundary points using the seed point as a reference point;
a rectangular search box obtaining module, configured to obtain a smallest bounding rectangle search box formed by the four initial road boundary points;
a first candidate search box obtaining module, configured to rotate the obtained search box at different angles, and move rotated search boxes to obtain a plurality of candidate search boxes;
a determining module, configured to determine whether a sum of squares of grayscale differences between the plurality of candidate search boxes and the obtained search box is greater than a preset threshold;
a road recognition module, configured to stop searching and complete road recognition when the sum of the squares of the grayscale differences between the plurality of candidate search boxes and the obtained search boxes is greater than the preset threshold; and
a second candidate search box obtaining module, configured to select a new search box from the plurality of candidate search boxes when the sum of the squares of the grayscale differences between the plurality of candidate search boxes and the obtained search boxes is not greater than the preset threshold, and retrieve a plurality of candidate search boxes based on the new search box.

### Claim 9 (depends on 8)

The road recognition system according to claim 8, wherein the search module comprises:
a first searching unit, configured to separately search for pixels in α° direction and (α+180)° direction of the seed point;
a determining unit, configured to determine whether a gray value difference between a current searched pixel and a previous pixel is greater than a preset threshold;
a second searching unit, configured to use the current pixel as one initial road boundary point when the gray value difference between the current searched pixel and the previous pixel is greater than the preset threshold, and then change the search direction with the seed point as the reference point to continue searching for a pixel; and
a third searching unit, configured to directly change the search direction with the seed point as the reference point to continue searching for a pixel when the gray value difference between the current searched pixel and the previous pixel is less than or equal to the preset threshold.

### Claim 10 (depends on 8)

The road recognition system according to claim 8, wherein the first candidate search box obtaining module comprises:
a rotating unit, configured to rotate the obtained search box with a center point of the obtained search box as a center; and
a candidate search box obtaining unit, configured to advance the rotated search box by a side length of the obtained search box to obtain candidate search boxes advanced at different rotation angles.
