Demonstrates enhanced defect classification in girth welds using YOLOv8, suggesting improved automated evaluation in engineering.
Key Points
The aim is to improve the accuracy and efficiency of defect identification in girth welds using advanced image processing and deep learning techniques.
Median filtering applied for denoising prior to defect analysis.
Spatial enrichment algorithm utilized to enhance image details.
YOLOv8 model implemented for the final defect recognition process.
Noise reduction achieved with PSNR values of around 36 for various defect types.
Spatial enrichment yielded a PSNR of 33.71 and an SSIM of 0.96.
The overall approach provided a superior balance between precision and recall.