Polarization Sensor Positioning for Object Recognition Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The recognition rate of objects in real space, including estimating position, posture, and shape, is hindered by suboptimal information acquisition conditions, leading to decreased accuracy in object recognition.

Innovation Solution

An information processing system employing a depth sensor and polarization sensor to acquire depth information and polarization information, which are then used to guide the user in positioning the polarization sensor for improved reliability and accuracy of object recognition, utilizing augmented reality technology to superimpose virtual objects on real objects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If polarization information is acquired under arbitrary acquisition conditions, then the information acquisition process is simple and fast, but the recognition accuracy and reliability of object recognition deteriorates

Engineering Contradiction:
Improverecognition accuracyVSAvoidinformation acquisition process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system pre-calculates and stores optimal acquisition conditions (拍摄条件) for polarization information based on different object types and recognition scenarios. Before actual recognition, the system selects appropriate pre-established conditions rather than performing real-time optimization, thereby improving recognition accuracy while avoiding complex real-time computation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system adjusts polarization information acquisition parameters (such as polarization angles, exposure time, illumination intensity) based on the selected acquisition conditions. By changing these parameters according to pre-established condition sets, the system optimizes recognition accuracy for different scenarios without requiring complex adaptive algorithms.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If multiple polarization images are acquired under different conditions to improve recognition accuracy, then the recognition reliability improves, but the information acquisition time and processing complexity increases

Engineering Contradiction:
Improverecognition reliabilityVSAvoidinformation acquisition time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system pre-establishes multiple sets of acquisition conditions corresponding to different object characteristics and recognition requirements. When recognition is needed, the system directly applies these pre-prepared condition sets, avoiding the need to acquire all possible polarization images under all conditions, thus balancing reliability with time efficiency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Instead of acquiring complete polarization information under all possible conditions, the system selectively acquires polarization images only under the specific pre-established conditions most relevant to the current recognition task. This partial action approach maintains sufficient recognition reliability while significantly reducing acquisition time.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If the polarization sensor is positioned arbitrarily, then the operation is simple and quick, but the quality and reliability of polarization information deteriorates

Engineering Contradiction:
Improvepolarization information qualityVSAvoidsensor positioning operation
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The system provides real-time feedback to the user regarding the current polarization sensor positioning quality. When the sensor is not in the optimal position according to the pre-established conditions, the system notifies the user and guides adjustment, ensuring high polarization information quality while maintaining ease of operation through intuitive feedback mechanisms.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system automatically adjusts or selects the optimal polarization sensor positioning based on pre-established conditions and current recognition requirements, reducing the need for manual positioning operations. The system serves itself by autonomously determining and implementing the correct sensor orientation for each recognition task.

Inventive Principle:
Principle #25Self-service

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

Enhances the accuracy of object recognition by optimizing the acquisition of polarization information, improving the reliability of the recognition process and resolving issues related to indefiniteness in polarization normals, thereby improving the overall recognition of objects in real space.

Implementation Method 1

a position, a posture, and a shape of an object in a real space on the basis of polarization information acquired by a polarization sensor

Methodology Applied
Scientific EffectPolarization: Polarisation

Implementation Method 2

depth information and polarization information acquired by a depth sensor and a polarization sensor, respectively

Methodology Applied
Scientific EffectTime of Flight: Time of Flight

Data Source

PatentEP3693925B1Information processing device, information processing method, and recording medium
Publication Date: 2022.10.26 SONY GROUP CORP
  • EP3693925B1 patent drawingFigure 1
  • EP3693925B1 patent drawingFigure 2
  • EP3693925B1 patent drawingFigure 3

AI summary

It is possible to acquire information used for recognizing an object in a real space in a more suitable manner. An information processing apparatus includes: an estimating unit that estimates a normal on at least a part of a face of an object in a real space on the basis of polarization information corresponding to a detection result of each of a plurality of beams of polarized light acquired by a polarization sensor and having different polarization directions; and a control unit that controls output of notification information for guiding a change in a position in the real space according to an estimation result of the normal.