Dual-Camera Relative Position Calculation via Image Screening

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for calculating dual-camera relative positions are inaccurate when the scene is difficult to recognize, such as when objects are far away, mostly planar, or lack texture, leading to inconsistent camera positioning due to environmental factors like drops or temperature changes.

Innovation Solution

A method involving the collection and screening of multiple dual-camera image groups to determine suitable images for calculating relative position parameters, using homography matrices and feature points to assess projection errors, and optimizing relative translation and rotation parameters through non-linear formulas and RANSAC algorithms to achieve accurate camera positioning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If current algorithms are used to calculate dual-camera relative position, then the calculation can be performed, but accuracy cannot be ensured when the scene is difficult to recognize (far objects, planar surfaces, or lack of texture)

Engineering Contradiction:
Improverelative position calculation accuracyVSAvoidscene recognition difficulty
Core Design Contradiction:
Measurement precisionVSDifficulty of detecting and measuring

Solution Approach 1:

The patent applies preliminary action by pre-screening multiple groups of dual-camera images before calculation, selecting only those groups that meet quality criteria (sufficient feature points, appropriate projection errors). This preliminary selection ensures that the subsequent relative position calculation is performed on high-quality images, thereby guaranteeing accuracy even when individual scenes are difficult to recognize.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes parameters by evaluating multiple image groups with different characteristics and selecting the optimal group based on quantitative criteria such as feature point quantity, projection error metrics, and scene complexity. By varying the input image selection parameters rather than using a fixed algorithm on any image, the system adapts to different scene conditions and maintains calculation accuracy.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If multiple groups of dual-camera images are screened and processed, then accurate relative position can be obtained, but the calculation complexity and processing time increase

Engineering Contradiction:
Improverelative position calculation accuracyVSAvoidcalculation process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent reduces calculation complexity through preliminary action by pre-evaluating multiple image groups and selecting only the most suitable ones for calculation. By performing quality assessment beforehand (checking feature point density, projection errors, scene characteristics), the system avoids performing complex calculations on unsuitable images, thereby reducing overall computational burden while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent applies taking out by extracting and selecting only the essential and most suitable image groups from the complete set of captured images. Rather than processing all captured images, the system extracts those that meet specific quality criteria, thereby simplifying the calculation process while ensuring accuracy through selective processing of high-quality data.

Inventive Principle:
Principle #2Taking out (Extraction)

3Reliability

If the relative position of dual cameras is recalculated frequently, then positioning accuracy is maintained, but processing time and computational resources are consumed

Engineering Contradiction:
Improvecamera positioning consistencyVSAvoidrecalculation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-screening and pre-evaluating multiple image groups before actual calculation, selecting only the most suitable groups for processing. This preliminary preparation ensures that when recalculation is needed, the system can quickly identify and process appropriate images, reducing the time penalty associated with frequent recalculations while maintaining positioning consistency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent applies partial action by processing only a selected subset of captured images rather than all images. By evaluating multiple image groups and selecting only those that meet quality criteria, the system performs calculations on a partial set of images, thereby reducing processing time and computational resource consumption while still achieving reliable positioning results.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3252715B1Two-camera relative position calculation system, device and apparatus
Publication Date: 2020.12.23 HUAWEI TECH CO LTD
  • EP3252715B1 patent drawingFigure 1
  • EP3252715B1 patent drawingFigure 2
  • EP3252715B1 patent drawingFigure 3

AI summary

The present invention provides a method and an apparatus for calculating a dual-camera relative position, and a device. The method for calculating a dual-camera relative position provided in the present invention includes: obtaining M groups of dual-camera images, where each group of dual-camera images in the M groups includes two images that are obtained by photographing a same scene by dual cameras at the same time; screening the M groups of dual-camera images, to obtain Q groups of dual-camera images that can be used to calculate the dual-camera relative position; obtaining dual-camera relative position parameters corresponding to each group of dual-camera images in the Q groups according to each group of dual-camera images in the Q groups; and obtaining optimal dual-camera relative position parameters from the Q groups of dual-camera relative position parameters. In the present invention, the dual-camera relative position can be calculated when a scene is relatively difficult to recognize.