Method for quantitative evaluation of metal powder
The method addresses the limitations of existing image-based evaluation by using AI segmentation to create a 3D polygon image from 2D slices, ensuring robust and precise assessment of metal powder properties and gas pores.
Patent Information
- Application Number
- JP2024017268
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-02-07
- Publication Date
- 2025-08-20
AI Technical Summary
Existing methods for quantitatively evaluating metal powders used in additive manufacturing rely on image processing parameters, which lack robustness and fail to account for three-dimensional continuity, making accurate evaluation of powder particles and gas pores challenging.
A method involving slicing a metal powder layer to obtain two-dimensional images, segmenting powder regions using an AI model, connecting these regions in a stacking direction to form a voxel 3D image, and converting it into a polygon 3D image for quantitative evaluation, independent of image processing parameters.
Enables stable and accurate quantitative evaluation of metal powder particles and gas pores without relying on image processing parameters, providing comprehensive three-dimensional analysis.
Smart Images

Figure 2025121673000001_ABST
Abstract
Citation Information
Patent Citations
Aluminum-based metal powder and method for producing same
JP2023503778A