Object Posture Estimation Using Depth Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing object grasping systems fail to accurately estimate the posture of objects, particularly those with complex shapes or uneven surfaces, leading to potential misgrasping by robots.

Innovation Solution

A method that involves imaging an area with multiple objects, creating small areas around each object, estimating the object type, selecting the most reliable area, estimating the depth of the selected area, and finally estimating the three-dimensional posture of the object for precise robotic grasping.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If only two-dimensional image processing is used to estimate object posture, then the processing complexity is reduced, but the accuracy of posture estimation for objects with complex shapes or uneven surfaces deteriorates

Engineering Contradiction:
Improveprocessing complexityVSAvoidposture estimation accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent transitions from two-dimensional image processing to three-dimensional depth estimation by introducing depth information. The depth estimation unit calculates depth values for multiple points on the object surface based on the two-dimensional image, thereby adding the depth dimension to recover three-dimensional posture information that cannot be obtained from two-dimensional images alone.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent segments the object into multiple points distributed across its surface. The depth estimation unit calculates depth values for these discrete points, and the posture estimation unit uses these segmented depth information to determine the overall posture. This segmentation allows complex three-dimensional posture estimation while maintaining computational feasibility.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If depth information is added to posture estimation, then the accuracy of grasping objects with complex geometries is improved, but the processing time and computational load increase

Engineering Contradiction:
Improvegrasping accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The depth estimation unit performs preliminary calculations by estimating depth values for multiple points on the object surface before the posture estimation step. This preliminary depth estimation prepares the necessary three-dimensional information in advance, allowing the posture estimation unit to efficiently determine the object's orientation and position without redundant computations.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent estimates depth for multiple discrete points on the object surface rather than computing complete three-dimensional models. This partial depth estimation approach provides sufficient information for accurate posture determination while significantly reducing computational complexity compared to full three-dimensional reconstruction methods.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20250073910A1Method For Estimating Posture Of Object, Control Device, And Robot System
Publication Date: 2025.03.06 SEIKO EPSON CORP
  • US20250073910A1 patent drawing
  • US20250073910A1 patent drawing
  • US20250073910A1 patent drawing

AI summary

A method for estimating a posture of an object includes: an image acquisition step of acquiring, by an imaging unit configured to image an area in which a plurality of objects are arranged, a two-dimensional image including at least one of the plurality of objects; a small area creation step of creating, for at least one object among the at least one object in the two-dimensional image, a small area surrounding the object; a type estimation step of estimating a type of the object surrounded by the small area; a selection step of selecting one of the small areas in which the type of the object has been estimated in the type estimation step; a depth estimation step of estimating a depth of the one of the small areas based on the two-dimensional image; and a posture estimation step of estimating a three-dimensional posture of the object included in the one of the small areas based on the type of the object and the depth estimated in depth estimation step.