Accumulated and aggregated shifting of intensity for defect detection on micro 3D textured surfaces

Research Progress of Automated Visual Surface Defect Detection for Industrial Metal Planar Materials

“…In order to be self-adaptive to the shape, size, and scale of the defects, Zhou et al [ 83 ] put forward a surface defect detection model based on double low-rank and sparse decomposition to detect defects as the saliency part of the image, which is not only robust to noise and illumination inhomogeneity but also highly adaptive to the complex and changeable surface defects of a steel plate. After the model was tested on the Northeastern University (NEU) surface defect dataset, the F measure was 0.606, and the calculation time on each image was 0.713 s. Yan et al [ 95 ] developed a probabilistic saliency framework based on a feature enhancement mechanism for realizing robust defect detection on a micro 3D texture surface of industrial products, which designed the absolute strength deviation and local strength aggregation to represent the initial saliency of the pixel level while all pixels are classified as defective or non-defective. To address the issues of intra-class defects having large differences in appearance while inter-class defects contain similar parts, Song et al have studied many approaches to combine visual salience with other ideas, such as Encoder–Decoder Residual network (EDRNet) [ 96 ], multiple constraints and improve texture feature (MCITF) [ 97 ], attention mechanism [ 98 ], and pyramid feature (PGA-Net) [ 99 ], and the experimental results show that they are both effective and outperform the state-of-the-art methods.…”

A Weakly Supervised Surface Defect Detection Based on Convolutional Neural Network

“…Each pixel is classified as a defect or non-defect, based on features calculated from neighboring pixels. The latest research work [14] reports that a novel probabilistic salience framework is based on two saliency features (absolute intensity deviation and local intensity aggregation) and is proposed to shift the intensity of each pixel according to its saliency during its iterative process. Common features involve geometric and statistical descriptors (i.e., length, width, area, mean, and standard deviation), as well as localized wavelet decomposition [15].…”

Robust defect detection in 2D images printed on 3D micro‐textured surfaces by multiple paired pixel consistency in orientation codes

“…The changes in illumination on such surfaces have a significant influence on the appearance, which causes difficulties in defect detection. We proposed another method for defect detection in background regions, called an accumulated and aggregated shifting of intensity, in short AASI [25], for the same type of surfaces. In this study, we concentrate on a different problem of defect detection on printed logotypes on the same objects.…”