Stereo Image Positioning via Segmented Depth Computation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing stereo image techniques require high-resolution images and sophisticated computation for accurate positional information acquisition, leading to increased processing load and a trade-off between accuracy and speed, especially when updating positional information at frame rates for motion-based processing.

Innovation Solution

The system divides positional information generation into two blocks: a first block for approximate position detection and a second block for detailed position acquisition, restricting search ranges and using high-resolution images and sophisticated algorithms only where necessary, allowing for efficient and accurate positional information processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If high-resolution stereo images and sophisticated computation techniques are used to accurately acquire positional information, then measurement precision is improved, but processing load and computation time increase significantly

Engineering Contradiction:
Improvepositional information accuracyVSAvoidprocessing speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The positional information acquisition process is divided into two distinct blocks: a first block that performs approximate position detection using simplified methods, and a second block that performs detailed position acquisition using high-resolution images and sophisticated algorithms. This segmentation allows the system to achieve high measurement precision for critical regions while maintaining overall processing speed by avoiding expensive computations across the entire image space.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system applies different processing qualities to different regions of the image based on their importance. The second positional information acquisition block focuses computational resources on specific regions of interest where high precision is required, while using simpler methods in the first block for general approximate positioning. This local quality approach optimizes the balance between measurement precision and processing efficiency.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If sophisticated computation techniques are applied to detect corresponding points in high-resolution stereo images, then measurement precision is improved, but processing time increases

Engineering Contradiction:
Improvecorresponding point detection accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The corresponding point detection process is segmented into two stages: first, approximate corresponding points are identified using simplified computation on lower-resolution images; second, only these approximate points serve as seeds for refined detection in high-resolution images. This segmentation dramatically reduces the time required for sophisticated computation by limiting its application to specific regions rather than the entire image.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The first positional information acquisition block performs preliminary corresponding point detection using simplified methods before the second block applies sophisticated algorithms. This preliminary action establishes initial estimates that guide the more computationally intensive processing, reducing the search space and time required for high-precision corresponding point detection.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If detailed stereo matching is performed across the entire image, then measurement precision is improved, but device complexity and processing load increase

Engineering Contradiction:
Improvepositional information accuracyVSAvoidprocessing system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The stereo matching process is divided into two functional blocks with different complexity levels. The first block uses simplified stereo matching algorithms suitable for rapid processing, while the second block applies detailed stereo matching only where necessary. This segmentation reduces overall device complexity by avoiding the deployment of maximum-complexity algorithms throughout the entire processing pipeline.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of applying detailed stereo matching to the entire image (excessive action), the system applies it only to specific regions or points where high precision is required (partial action). This partial application of the complex algorithm reduces processing load and system complexity while maintaining measurement precision for critical tasks.

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 high-speed and high-accuracy information processing of object positions and motions, reducing processing load while maintaining accuracy, by using simplified methods for initial approximation and detailed processing only where needed.

Implementation Method 1

an imaging device for imaging an object

Methodology Applied
Scientific EffectLight reflection: Reflection

Implementation Method 2

corresponding points are detected from stereo images of a same space simultaneously taken with two cameras horizontally separated from each other by a known interval and, on the basis of a resultant parallax between the detected points, a distance from an imaged surface of an object is computed by use of the principle of triangulation

Methodology Applied
Scientific EffectParallax: Parallax

Implementation Method 3

a distance from an imaged surface of an object is computed by use of the principle of triangulation

Methodology Applied
Scientific EffectTriangulation:

Data Source

PatentEP3343500B1Image processing device and image processing method
Publication Date: 2021.06.30 SONY INTERACTIVE ENTERTAINMENT LLC
  • EP3343500B1 patent drawingFigure 1~2
  • EP3343500B1 patent drawingFigure 3~4
  • EP3343500B1 patent drawingFigure 5

AI summary

An image acquisition block 42 of an information processing apparatus 10 acquires stereo images taken by a first camera 13a and a second camera 13b that make up an imaging apparatus 12. An input information acquisition block 44 accepts a user manipulation. A first positional information acquisition block 52 of a positional information generation block 46 identifies an approximate position of an object to be imaged through predetermined means. A second positional information acquisition block 56 determines an estimated distance range of a target on the basis of the identified approximate position of the object to be imaged, detects corresponding points by executing block matching on stereo images thorough only a search range corresponding to the determined estimated distance range, and obtains a position of the target with high resolution and accuracy. An output information generation block 50 generates output data on the basis of the position of that target and outputs the generated output data.