3D Surface Inspection Path Planning for Sharp Image Acquisition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robotic control systems using optical matrix sensors struggle to accurately control surfaces with complex geometry in industrial settings, as they require precise positioning and trajectory design that is challenging to achieve with traditional experimental learning methods.

Innovation Solution

The method involves generating a three-dimensional virtual model of the sharpness volume, which includes loading a three-dimensional model of the device, generating a truncated pyramid representing the optics, and tiling the surface with unit models of the sharpness volume to define acquisition positions and trajectories for the robot.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional experimental learning methods are used to design robot trajectories, then the robot can be trained directly on the part, but the process is time-consuming and difficult to achieve precise positioning for complex geometry surfaces

Engineering Contradiction:
Improvepositioning precisionVSAvoidtrajectory design time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent creates a three-dimensional virtual model of the part and simulates the robot trajectory in advance before actual execution. This preliminary virtual modeling and simulation allows the system to plan the optimal path without time-consuming experimental learning, achieving precise positioning for complex surfaces through pre-computed waypoints and trajectory parameters.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates a virtual copy of the physical part in three-dimensional space. This digital replica allows trajectory design and optimization to be performed in the virtual environment, eliminating the need for repeated experimental learning on the actual part and significantly reducing setup time while maintaining positioning precision.

Inventive Principle:
Principle #26Copying

2Area of stationary object

If multiple acquisitions from different viewpoints are performed on large parts, then complete surface coverage is achieved, but the number of acquisitions increases and the inspection time extends

Engineering Contradiction:
Improvesurface coverage areaVSAvoidinspection speed
Core Design Contradiction:
Area of stationary objectVSProductivity

Solution Approach 1:

The patent transitions from two-dimensional image acquisition to three-dimensional virtual modeling. By creating a comprehensive 3D virtual model that includes depth and spatial relationships, the system can plan optimal viewpoints and trajectories in three-dimensional space, reducing the number of acquisitions needed to achieve complete surface coverage while maintaining inspection quality.

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

Solution Approach 2:

The patent pre-calculates optimal acquisition viewpoints and robot trajectories in the virtual environment before execution. This preliminary planning identifies the minimum number of viewpoints required to achieve complete surface coverage, eliminating redundant acquisitions and accelerating the inspection process.

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If the robot trajectory is designed by experimental learning directly on the part, then the trajectory can be adapted to the actual part geometry, but the process requires physical access to the part and extended setup time

Engineering Contradiction:
Improvetrajectory adaptabilityVSAvoidtrajectory setup ease
Core Design Contradiction:
Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The patent uses a three-dimensional virtual model as a digital copy of the physical part. This virtual replica preserves all geometric features and surface characteristics, allowing the robot trajectory to be adapted to the actual part geometry through accurate digital representation. The virtual model enables trajectory design without physical access to the part, significantly simplifying setup while maintaining adaptability.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent performs trajectory adaptation in advance within the virtual environment. By pre-adjusting the trajectory to match the digital model's geometry, the system achieves part-specific optimization without requiring time-consuming experimental learning on the actual part, making the process more efficient and easier to implement.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3658998B1Surface inspection method
Publication Date: 2025.06.18 SAFRAN SA
  • EP3658998B1 patent drawingFigure 1~3
  • EP3658998B1 patent drawingFigure 4~5
  • EP3658998B1 patent drawingFigure 6~7

AI summary

The invention relates to a method for inspecting a surface (1) of a workpiece (2) using an image capture device to be mounted on a robot (4), said image capture device (3) comprising a sensor and a lens associated with: an optical centre C, an angular aperture (alpha), and a depth of field (DoF), and defining a sharpness volume (6). The method comprises the following operations: loading a virtual three-dimensional model of the surface (1); generating a virtual three-dimensional model of the sharpness volume (6); tiling the model of the surface (1) using a plurality of unit models of the virtual three-dimensional model of the sharpness volume (6); for each position of the unit models (6), calculating the corresponding position, known as the acquisition position, of the image capture device (3).