Automated Detection of Micro-Scale Porosity Defects in Reflective Metal Parts via Deep Learning and Polarization...

Area of Science:

• Materials Science
• Computer Vision
• Manufacturing Engineering

Background:

• Defect detection in precision additive manufacturing of highly reflective metals is challenging.
• Existing methods struggle with small micro and nano-scale defects on these surfaces.

Purpose of the Study:

• To enhance intelligent identification of small metal micro and nano defects.
• To improve defect detection accuracy in additive manufacturing of high-precision metal materials.

Main Methods:

• Proposed an enhanced SCK-YOLOV5 framework combining polarization imaging and deep learning.
• Introduced a novel SNWD (Selective Network with attention for Defect and Weathering Degradation) Loss function.
• Employed global space construction with dual-attention and multi-scale feature refining using selection kernel convolution.

Main Results:

• Significantly improved precision, recall rate, and mAP50 index compared to the YOLOv5 baseline.
• Achieved a 0.5% increase in precision, 1.2% in recall, and 1.8% in mAP50.
• Demonstrated stable extraction of multi-scale defect information from highly reflective surfaces.

Conclusions:

• The enhanced SCK-YOLOV5 framework offers a new paradigm for intelligent defect detection in additive manufacturing.
• Provides reliable technical support for quality control in industrial manufacturing of high-precision metal materials.
• Represents a novel advancement in YOLO-based defect detection methods.