Lane Image Thresholding for Solid and Dotted Line Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing lane detection methods in autonomous vehicles face challenges in accurately distinguishing between solid and dotted lines while suppressing noise, often leading to the loss of important feature points and picking up unnecessary noise points due to the use of a single threshold.

Innovation Solution

An image processing method that divides the difference image into multiple subareas based on the center line, determining separate thresholds for solid and dotted lines using statistical analysis to create adapted thresholds for each subarea, thereby enhancing the detection of lane features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a lower threshold is used to detect edges of dotted lines, then more edges can be detected, but a large number of noise points are picked up

Engineering Contradiction:
Improveedge detection accuracyVSAvoidnoise points
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The image processing area is divided into multiple subareas (first subarea for solid lines, second subarea for dotted lines). Each subarea has its own threshold value determined through statistical analysis of pixel gray values specific to that subarea. This segmentation allows the system to detect dotted line edges without picking up excessive noise, while simultaneously detecting solid lines effectively.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different threshold values are applied to different subareas based on their local characteristics. The first threshold value is determined for the first subarea (solid line region) and the second threshold value is determined for the second subarea (dotted line region). This local adaptation of threshold quality enables optimal detection for each line type without compromising the other.

Inventive Principle:
Principle #3Local quality

2Object-affected harmful factors

If a higher threshold is used to suppress noise, then fewer noise points are picked up, but edges of dotted lines may be missed

Engineering Contradiction:
Improvenoise pointsVSAvoidedge detection accuracy
Core Design Contradiction:
Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The image is segmented into distinct subareas for solid and dotted lines, allowing separate threshold optimization. The second subarea dedicated to dotted lines uses a threshold specifically adapted to its characteristics, ensuring dotted line edges are not missed while the first subarea handles solid lines with its own threshold.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Each subarea receives a locally optimized threshold value based on statistical analysis of its specific pixel distribution. This local quality adaptation ensures that the threshold is appropriate for the local line type (solid or dotted) without being overly restrictive or permissive.

Inventive Principle:
Principle #3Local quality

3Device complexity

If a single threshold is used for the entire image, then the processing is simple, but important feature points are lost and noise cannot be suppressed

Engineering Contradiction:
Improveprocessing complexityVSAvoidfeature point detection accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

Rather than using a single threshold for the entire image, the method segments the image into multiple subareas and determines separate threshold values for each subarea through statistical analysis. This segmentation approach maintains reasonable processing complexity while dramatically improving feature point detection accuracy and noise suppression.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The threshold parameter is changed from a single global value to multiple local values, with each threshold determined through statistical analysis of the specific subarea's pixel gray value distribution. This parameter adaptation allows the system to respond to local image characteristics while maintaining efficient processing.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3872690B1Image processing method and apparatus used for lane detection
Publication Date: 2026.04.22 GREAT WALL MOTOR CO LTD
  • EP3872690B1 patent drawingFigure 1
  • EP3872690B1 patent drawingFigure 2
  • EP3872690B1 patent drawingFigure 3(a)~3(d)

AI summary

The invention relates to the field of intelligent transportation and image processing, and provides an image processing method and apparatus for lane detection. The image processing method according to the invention includes: acquiring a road image captured by a camera on a vehicle; determining a rectangular area covering a lane in the road image; carrying out difference processing on the rectangular area to acquire a corresponding difference image; dividing the difference image into a plurality of subareas independently comprising a solid line or a dotted line of the lane; determining a threshold adapted to pixels of each subarea; and processing the difference image of the corresponding subarea based on the determined threshold to acquire a corresponding ternary image. The invention determines different thresholds for different areas including solid lines or dotted lines through a multi-area multi-threshold mode, and then acquires corresponding ternary images based on the determined thresholds, thus solving a problem that the solid lines and the dotted lines are difficult to pick up through a single threshold in transfer lane detection, avoiding loss of important feature points and being beneficial for suppressing noise points.