Stereo Camera Parallax Correction via Region Segmentation

Solution Overview

Problem

Existing distance measurement apparatuses installed in vehicle cabins face reduced detection accuracy due to windshield distortion, affecting the measurement of distances to objects ahead of the vehicle.

Innovation Solution

A distance measuring device with an image acquiring unit, parallax detection unit, and error correction mechanisms that identify and correct parallax errors by dividing the imaging region into correction regions, allowing for individual parallax adjustments to improve measurement accuracy despite windshield distortion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If two imaging devices are installed in the vehicle cabin to capture images ahead of the vehicle, then distance measurement can be performed using parallax detection, but measurement accuracy is reduced due to windshield distortion

Engineering Contradiction:

Improvedistance measurement accuracyVSAvoidparallax detection accuracy

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The imaging region is divided into multiple correction regions (first correction region, second correction region, etc.), each with its own correction coefficient. This segmentation allows different parts of the image to be corrected using region-specific coefficients, thereby improving distance measurement accuracy while accounting for windshield distortion variations across different fields of view.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different correction coefficients are applied to different correction regions based on their specific characteristics. The first correction coefficient is applied to the first correction region while the second correction coefficient is applied to the second correction region, allowing each region to be corrected according to its local distortion characteristics rather than using a uniform correction approach.

Inventive Principle:

Principle #3Local quality

2Device complexity

If a single correction coefficient is used for the entire imaging region, then correction process is simple, but parallax errors vary across different regions due to windshield distortion

Engineering Contradiction:

Improvecorrection process complexityVSAvoidparallax detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The imaging region is divided into multiple correction regions (first correction region, second correction region, etc.), each with its own correction coefficient. This segmentation allows different parts of the image to be corrected using region-specific coefficients, thereby improving distance measurement accuracy while accounting for windshield distortion variations across different fields of view.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different correction coefficients are applied to different correction regions based on their specific characteristics. The first correction coefficient is applied to the first correction region while the second correction coefficient is applied to the second correction region, allowing each region to be corrected according to its local distortion characteristics rather than using a uniform correction approach.

Inventive Principle:

Principle #3Local quality

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The device enhances measurement accuracy by correcting parallax errors in each correction region, ensuring precise distance calculations even with varying parallax errors caused by windshield distortion, without requiring dedicated calibration for the stereo camera.

Implementation Method 1

acquiring a plurality of images which are simultaneously captured from mutually different view points via a windshield of the vehicle

Methodology Applied

Scientific EffectLight reflection: Reflection

Implementation Method 2

a parallax of the object present in the two images is detected based on the two captured images, thereby detecting the distance up to the object

Methodology Applied

Scientific EffectParallax: Parallax

Implementation Method 3

The distance detection unit is mounted on the vehicle, transmitting/receiving electromagnetic waves to detect distance between the front object and the vehicle as electromagnetic waves detected distance

Methodology Applied

Scientific EffectElectromagnetic wave transmission and reception: Radar

Data Source

PatentUS10338211B2Apparatus for measuring distance

Publication Date: 2019.07.02 DENSO CORP

AI summary

The distance measuring device detects parallax between right and left captured images where a front object is present ahead of the own vehicle, and calculate a parallax distance between the front object and the own vehicle. The radar device detects a radar detected distance therebetween. The distance measuring device calculates a parallax error based on the difference between the parallax calculated distance and the radar detected distance. A correction region is set in the right captured image and divided into correction divided regions. The distance measuring device sets an individual correction parallax for every correction divided regions, based on the parallax error and the identified correction divided regions.