Digital Image Processing
An object detection and recognition system for weld bead extraction from digital radiographs
Computer Vision and Image Understanding
Hi-index | 0.00 |
A new approach for defects detection in low contrast digital images (X-ray or photos) and images with uneven background illumination is presented in this paper. The algorithm comprises two main stages: image pre-processing (noise suppression and correction of the uneven illumination) and adaptive defects segmentation. This approach permits the successful detection of different kinds of defects and irregularities and ensures high accuracy. It is suitable for the analysis of X-ray images of welds, or photos of pipes, plates, etc. The experimental results obtained with the software implementation of the described algorithms prove their efficiency. The paper also points up the advantages of the presented algorithm in comparison with some well-known methods for non-destructive control.