Lane Detection via Distorted Trapezoidal ROI Merging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lane detection systems face challenges in accurately detecting lanes under unfavorable conditions, such as complex road environments, where the movement of vehicles and distorted lane markings complicate the detection process, leading to reduced detection performance and accuracy.
Innovation Solution
A device and method that utilize a first ROI setting unit to determine a rectangular ROI, a second ROI setting unit to create a distorted trapezoidal ROI by merging line segments and adjusting slopes, and a lane detection unit to detect lane markings, incorporating color difference information and brightness scanning to enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a simple rectangular ROI is used for lane detection, then the detection process is simple and fast, but detection accuracy deteriorates under unfavorable conditions such as distorted lane markings
Solution Approach 1:
The detection process is divided into two distinct phases: first extracting line segments from the rectangular ROI, then merging these segments to form a distorted trapezoidal ROI. This segmentation allows the system to maintain the simplicity of rectangular extraction while achieving the accuracy benefits of distorted shape adaptation.
Solution Approach 2:
Line segments are extracted and merged in advance to determine the distorted trapezoidal ROI before actual lane detection. This preliminary action prepares the optimal detection region beforehand, enabling accurate lane marking detection while maintaining processing efficiency through pre-computed geometric transformation.
2Measurement precision
If a distorted trapezoidal ROI is used to adapt to lane marking distortion, then detection accuracy improves, but the complexity of ROI setting increases
Solution Approach 1:
The complex distorted trapezoidal ROI is constructed by segmenting and merging simpler line segments. Instead of directly computing a complex distorted shape, the system breaks down the process into extracting multiple line segments and merging them, thereby reducing the complexity of individual operations while achieving the desired accurate ROI shape.
Solution Approach 2:
Multiple extracted line segments are merged to form the distorted trapezoidal ROI. This merging process combines simpler geometric elements into a more accurate representation of the lane marking region, achieving high detection accuracy without requiring direct computation of complex distorted geometries.
3Device complexity
If traditional edge detection methods are used, then the algorithm is simple, but detection performance deteriorates in complicated road environments such as boundary lanes or city areas
Solution Approach 1:
The detection algorithm is segmented into multiple specialized steps: line segment extraction, merging to form distorted ROI, and lane detection within the optimized region. This segmentation allows each step to be optimized for its specific function, improving reliability in complex environments while maintaining overall algorithmic simplicity through modular design.
Solution Approach 2:
The algorithm applies different processing strategies to different regions: using line segment extraction and merging for ROI determination, then applying lane detection specifically within the distorted trapezoidal ROI. This local quality approach tailors the detection method to the specific characteristics of each region, enhancing performance in complicated road environments.
Data Source
AI summary
A lane detection and lane departure determining device and method are provided. The lane detection device includes a first Region of Interest (ROI) setting unit configured to set a rectangular ROI from an input image which is a road image, a second ROI setting unit configured to set a Λ-ROI having distorted trapezoidal shape from the rectangular ROI, and a lane detection unit configured to detect the lane including a left and right lane marking in the Λ-ROI. The second ROI setting unit is configured to calculate a merged first line segment through a merging process, obtains a left and right second line segment by scanning brightness, and determines the left and right second line segment as a temporary left and right lane marking when a preset condition is satisfied. The Λ-ROI is determined by changing a slope of the temporary left and right lane marking in a predetermined range.