Additive Manufacturing Quality Assessment Using Binary Image Erosion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current quality assessment methods for additive manufacturing components, particularly those using generative laser methods, face challenges in accurately evaluating component quality due to misinterpretations from image data with varying brightness values and transitions between components and powder, leading to erroneous quality judgments.

Innovation Solution

A method involving digitization of image data, conversion to binary images, erosion to exclude transitions between components and powder, determination of contour data to define regions of interest, and classification based on the presence of quality defects, ensuring reliable and automatic quality assessment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If optical tomography is used to monitor layer structure, then image data for quality assessment is obtained, but misinterpretations occur due to transitions between component and powder causing erroneous quality judgments

Engineering Contradiction:
Improvequality assessment accuracyVSAvoidquality judgment reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The image data is segmented into two distinct parts: the component area and the powder area. By applying binary thresholding, the method separates pixels representing the component from those representing surrounding powder. This segmentation allows quality assessment to focus exclusively on the component area, eliminating misinterpretations caused by powder transitions and improving both measurement precision and judgment reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The method extracts the component area from the complete image by using binary thresholding and erosion operations. The extracted component mask isolates only the relevant region for quality assessment, removing the distracting powder areas that cause erroneous judgments. This extraction ensures that quality metrics are calculated solely from component pixels, enhancing assessment accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

2Loss of information

If brightness values are used for quality assessment, then information about component structure is obtained, but varying brightness values lead to misinterpretations and erroneous quality judgments

Engineering Contradiction:
Improvecomponent structure informationVSAvoidquality assessment accuracy
Core Design Contradiction:
Loss of informationVSMeasurement precision

Solution Approach 1:

The method transforms the brightness value parameter into a binary parameter through thresholding. Instead of analyzing varying brightness values that cause misinterpretation, the method converts pixel values into binary states (component vs. non-component). This parameter transformation preserves structural information while eliminating the ambiguity of varying brightness, thereby improving measurement precision without losing essential component structure data.

Inventive Principle:
Principle #35Parameter changes

3Area of stationary object

If complete image data including powder areas is assessed, then comprehensive coverage is achieved, but false quality deficiencies are detected at component edges

Engineering Contradiction:
Improveimage assessment coverageVSAvoidquality defect detection reliability
Core Design Contradiction:
Area of stationary objectVSReliability

Solution Approach 1:

The assessment area is segmented into component-specific regions and excluded powder regions. By applying binary thresholding and erosion, the method creates a precise component mask that defines the boundaries of the component area. Quality assessment is then performed only within this segmented component area, ensuring comprehensive coverage of the component while automatically excluding powder areas that would cause false defect detections at edges.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3070674B1Method of judging quality of a component that is produced by means of an additive manufacturing process
Publication Date: 2021.05.26 MTU AERO ENGINES GMBH
  • EP3070674B1 patent drawingFigure 1~2
  • EP3070674B1 patent drawingFigure 3~4

AI summary

The invention relates to a method for quality assessment of a component manufactured by an additive manufacturing process. The method includes at least the following steps: providing image data that characterizes at least one location of the manufactured component; converting the image data into a binary image (18); eroding the binary image (18) into a structure image (20); determining contour data (22) of the structure image (20); determining at least one image section (24) of the image data bounded by the contour data (22); checking the at least one image section (24) for the presence of an image area corresponding to a quality defect; and classifying the component as qualitatively acceptable if no quality defect is present, or classifying the component as qualitatively unacceptable if a quality defect is present. The invention further relates to a device for carrying out such a method.