Inspired by the coarse-to-fine visual perception process of human vision system,a new approach based on Gaussian multi-scale space for defect detection of industrial products was proposed.By selecting different scale ...Inspired by the coarse-to-fine visual perception process of human vision system,a new approach based on Gaussian multi-scale space for defect detection of industrial products was proposed.By selecting different scale parameters of the Gaussian kernel,the multi-scale representation of the original image data could be obtained and used to constitute the multi- variate image,in which each channel could represent a perceptual observation of the original image from different scales.The Multivariate Image Analysis (MIA) techniques were used to extract defect features information.The MIA combined Principal Component Analysis (PCA) to obtain the principal component scores of the multivariate test image.The Q-statistic image, derived from the residuals after the extraction of the first principal component score and noise,could be used to efficiently reveal the surface defects with an appropriate threshold value decided by training images.Experimental results show that the proposed method performs better than the gray histogram-based method.It has less sensitivity to the inhomogeneous of illumination,and has more robustness and reliability of defect detection with lower pseudo reject rate.展开更多
针对工业缺陷对比度低、周围干扰信息多导致的误检率和漏检率高的问题,提出一种基于改进YOLOv8的工业表面缺陷检测算法EML-YOLO。通过设计一种高效大卷积模块(efficient large kernel,ELK),在保留空间信息的同时提供多尺度的特征表示,...针对工业缺陷对比度低、周围干扰信息多导致的误检率和漏检率高的问题,提出一种基于改进YOLOv8的工业表面缺陷检测算法EML-YOLO。通过设计一种高效大卷积模块(efficient large kernel,ELK),在保留空间信息的同时提供多尺度的特征表示,从而提高模型的特征提取能力;提出多支路并行的特征融合模块(multi-scale context module,MCM),使得模型能够获取丰富的特征信息和全局上下文信息;在Neck模块中通过特征压缩和精简来减少模型的参数量和计算量,让模型更适用于资源有限的工业场景。采用GC10-DET和DeepPCB两个工业表面缺陷数据集来验证改进的EML-YOLO算法的有效性。实验结果表明,在GC10-DET数据集和DeepPCB数据集上,检测准确率上分别提高了4.3个百分点和2.9个百分点,参数量仅2.7×10^(6)。所提算法可以较好地应用于工业缺陷检测场景。展开更多
针对钢材表面缺陷纹理特征不明显、不同缺陷类间差异不明显和缺陷尺度变化剧烈等问题,本文设计了一个纹理信息增强模块(Texture Information Enhancement Module,TIEM)来保留主干网络上采样丢失的细节纹理特征信息和加强主干网络对不规...针对钢材表面缺陷纹理特征不明显、不同缺陷类间差异不明显和缺陷尺度变化剧烈等问题,本文设计了一个纹理信息增强模块(Texture Information Enhancement Module,TIEM)来保留主干网络上采样丢失的细节纹理特征信息和加强主干网络对不规则缺陷的空间建模能力;在颈部网络融入跳跃连接的多尺度自适应卷积模块(Multi-scale Adaptive convolution mod-ule with Skip Connections,MASC)来增强网络对不同尺度缺陷目标的感知能力,进而增强小目标的细粒度特征和大目标的高层语义信息,增强检测器的全局感知能力。以YOLOv7为基线模型,在公开数据集NEU-DET上,改进后的模型比基线模型mAP_(50)和mAP_(50:95)分别提高了3.0%和2.1%,并优于现阶段其他主流目标检测器。展开更多
基金supported in part by the Natural Science Foundation of China (NSFC) (Grant No:50875240).
文摘Inspired by the coarse-to-fine visual perception process of human vision system,a new approach based on Gaussian multi-scale space for defect detection of industrial products was proposed.By selecting different scale parameters of the Gaussian kernel,the multi-scale representation of the original image data could be obtained and used to constitute the multi- variate image,in which each channel could represent a perceptual observation of the original image from different scales.The Multivariate Image Analysis (MIA) techniques were used to extract defect features information.The MIA combined Principal Component Analysis (PCA) to obtain the principal component scores of the multivariate test image.The Q-statistic image, derived from the residuals after the extraction of the first principal component score and noise,could be used to efficiently reveal the surface defects with an appropriate threshold value decided by training images.Experimental results show that the proposed method performs better than the gray histogram-based method.It has less sensitivity to the inhomogeneous of illumination,and has more robustness and reliability of defect detection with lower pseudo reject rate.
文摘针对工业缺陷对比度低、周围干扰信息多导致的误检率和漏检率高的问题,提出一种基于改进YOLOv8的工业表面缺陷检测算法EML-YOLO。通过设计一种高效大卷积模块(efficient large kernel,ELK),在保留空间信息的同时提供多尺度的特征表示,从而提高模型的特征提取能力;提出多支路并行的特征融合模块(multi-scale context module,MCM),使得模型能够获取丰富的特征信息和全局上下文信息;在Neck模块中通过特征压缩和精简来减少模型的参数量和计算量,让模型更适用于资源有限的工业场景。采用GC10-DET和DeepPCB两个工业表面缺陷数据集来验证改进的EML-YOLO算法的有效性。实验结果表明,在GC10-DET数据集和DeepPCB数据集上,检测准确率上分别提高了4.3个百分点和2.9个百分点,参数量仅2.7×10^(6)。所提算法可以较好地应用于工业缺陷检测场景。
文摘针对钢材表面缺陷纹理特征不明显、不同缺陷类间差异不明显和缺陷尺度变化剧烈等问题,本文设计了一个纹理信息增强模块(Texture Information Enhancement Module,TIEM)来保留主干网络上采样丢失的细节纹理特征信息和加强主干网络对不规则缺陷的空间建模能力;在颈部网络融入跳跃连接的多尺度自适应卷积模块(Multi-scale Adaptive convolution mod-ule with Skip Connections,MASC)来增强网络对不同尺度缺陷目标的感知能力,进而增强小目标的细粒度特征和大目标的高层语义信息,增强检测器的全局感知能力。以YOLOv7为基线模型,在公开数据集NEU-DET上,改进后的模型比基线模型mAP_(50)和mAP_(50:95)分别提高了3.0%和2.1%,并优于现阶段其他主流目标检测器。
