Stereo Depth Map Generation via Adjustable Image Resolution Scaling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for generating depth maps with stereo cameras have a fixed minimum measurable distance and measurement resolution, limiting the flexibility and accuracy of depth measurement in various applications.

Innovation Solution

The proposed method involves an adjustable image resolution system that scales and processes stereo image pairs to generate depth maps with adjustable effective measurement ranges, allowing for flexible depth measurement by determining the appropriate image resolution based on the depth range of the target object.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If fixed-scale processing is used to generate depth maps, then the processing method is simple, but the minimum measurable distance and measurement resolution are fixed and cannot be adjusted

Engineering Contradiction:
Improveadjustability of measurement rangeVSAvoidprocessing complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies dynamics by making the image resolution adjustable rather than fixed. The system dynamically changes the resolution of stereo images before depth map generation, allowing the minimum measurable distance and measurement resolution to be adapted to different application scenarios. This resolves the contradiction by enabling measurement range adjustability while maintaining a relatively simple fixed-scale processing framework.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the resolution parameter of input images to control the measurement characteristics of the depth map. By adjusting the image resolution before processing, the system can control the minimum measurable distance and measurement resolution of the output depth map. This parameter change approach enables adaptability without fundamentally changing the processing methodology.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If high resolution is used for depth map generation, then the measurement precision is improved, but the processing time and computational load increase

Engineering Contradiction:
Improvedepth measurement resolutionVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by processing images at different resolutions selectively. Instead of always using the highest resolution, the system adjusts the image resolution to match the required measurement precision for each specific application. This allows obtaining sufficient measurement precision while avoiding the unnecessary computational load and processing time associated with always using maximum resolution.

Inventive Principle:
Principle #16Partial or excessive action

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

This approach enables depth maps with adjustable measurement resolutions, enhancing the accuracy and applicability of depth measurement in different scenarios, from short to long distances, by dynamically adjusting the image resolution to match the target depth range.

Implementation Method 1

A depth map may be extracted from a pair of images based on their image disparity. The pixel representing the largest image disparity may indicate the minimum measurable distance from the camera.

Methodology Applied
Scientific EffectParallax: Parallax

Data Source

PatentEP3391647B1Method, apparatus, and non-transitory computer readable medium for generating depth maps
Publication Date: 2023.11.01 BOE TECHNOLOGY GROUP CO LTD
  • EP3391647B1 patent drawingFigure 1~2
  • EP3391647B1 patent drawingFigure 3~4
  • EP3391647B1 patent drawingFigure 5~6

AI summary

A depth generation method is disclosed. The method includes obtaining a left two-dimensional (2D) image and a right two-dimensional image, each having a first image resolution; scaling the left 2D image and the right 2D image to obtain a scaled left 2D image and a scaled right 2D image, each having a second image resolution; and generating an output depth map based on the scaled left 2D image and the scaled right 2D image.