Guided Visual Inspection Navigation for Valid Camera Viewpoints

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing visual inspection systems face challenges in ensuring data quality and reliability due to limited flexibility and uncertainty in mobile data acquisition, particularly in semi-known environments, where aspects like occlusion, orientation, lighting, and asset dimensions impact the quality of collected data, leading to uncertain inspection conclusions.

Innovation Solution

A guided data acquisition method using 2D deep learning models for object recognition and navigation planning, which includes object localization, quality evaluation, and rule-based recommendations to adjust camera positions and navigation paths, ensuring that captured images meet predefined quality requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If mobile data acquisition systems are used to improve flexibility in inspection, then ease of operation is improved, but data quality control deteriorates due to uncertainty in localization and navigation

Engineering Contradiction:

ImproveflexibilityVSAvoiddata quality control

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system implements feedback mechanisms by continuously evaluating data quality metrics (sharpness, exposure, focus) during mobile acquisition and providing real-time guidance to operators or autonomous systems to adjust capture parameters, ensuring quality requirements are met despite mobility-induced uncertainties

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system dynamically adjusts acquisition parameters (shutter speed, ISO, focus distance, exposure compensation) based on real-time environmental conditions detected by sensors (lighting, distance, motion), allowing the mobile system to maintain data quality across varying operational conditions

Inventive Principle:

Principle #35Parameter changes

2Device complexity

If rule-based navigation planning is used to simplify navigation, then device complexity is reduced, but measurement precision deteriorates due to inability to adapt to quality requirements

Engineering Contradiction:

Improvenavigation planningVSAvoiddata quality

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The navigation planning system transitions from static rule-based paths to dynamic adaptive planning that continuously adjusts navigation trajectories based on real-time quality evaluation feedback, allowing the system to optimize capture positions and angles while maintaining manageable complexity through modular architecture

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs preliminary quality assessment and navigation planning by pre-calculating optimal capture positions and trajectories based on expected quality requirements, then adjusts during execution based on actual conditions, balancing advance preparation with adaptive response

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If SLAM and 3D-map reconstruction are used to improve localization accuracy, then measurement precision is improved, but device complexity and resource requirements increase

Engineering Contradiction:

Improvelocalization accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system extracts and uses only the essential localization information needed for navigation and quality assessment, avoiding full 3D-map reconstruction by focusing on key positional and orientational data required for capture planning, thereby reducing computational complexity while maintaining sufficient precision

Inventive Principle:

Principle #2Taking out (Extraction)

4Measurement precision

If deep learning models are used to improve object recognition and quality assessment, then measurement precision is improved, but use of energy and computational resources increase

Engineering Contradiction:

Improvequality assessment accuracyVSAvoidcomputational energy

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system applies deep learning models selectively only when needed for quality assessment and guidance generation, rather than continuously processing all data, and uses lightweight model variants optimized for edge devices, achieving sufficient precision while managing computational energy consumption

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system replaces complex mechanical or hardware-based quality control mechanisms with software-based deep learning assessment, enabling more precise and adaptive quality evaluation while reducing physical system complexity and enabling flexible deployment on various mobile platforms

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4040400B1Guided inspection with object recognition models and navigation planning

Publication Date: 2025.09.10 HITACHI LTD

Figure 1A~1C
Figure 2
Figure 3

AI summary

Example implementations involve systems and methods to advance data acquisition systems for automated visual inspection using a mobile camera infrastructure. The example implementations address the uncertainty of localization and navigation under semi-controlled environments. The approach combines object detection models and navigation planning to control the quality of visual inputs in the inspection process. The solution guides the operator (human or robot) to collect only valid viewpoints to achieve higher accuracy. Finally, the learning models and navigation planning are generalized to multiple type and size of inspection objects.