摘要
为应对实际工业产品视觉质量检测中缺陷罕见、尺寸小等挑战,提出了一种仅需要少量训练样本的小尺寸缺陷检测算法--Smooth-DETR,该算法采用基于DETR的编码-解码结构对缺陷类别和位置进行预测,该结构降低了参数量和计算复杂度。因DETR强大的全局特征学习能力,该算法可从少量训练样本中充分挖掘产品表面纹理特征,从而对打破了表面纹理连续性的缺陷检出率高;通过结合Smooth-L_(1)损失和GIoU损失的优势,进一步提升了小尺寸缺陷的回归精度。实验结果表明,所提方法检测性能优于现有先进检测模型。此外,仅用少量训练样本,该算法对11类产品表面的缺陷检测平均精确率就能够达到98%以上。
To deal with challenges of limited and small-size defects in product quality inspection,this paper proposed a method for surface-defect-detection of small-size with few training samples(Smooth-DETR).This method utilized DETR-based encoder-decoder to predict the classification and location of defects,which reduced the parameters and complexity.DETR had a strong global feature learning capability,which could obtain rich texture features of product surfaces with few samples,so that it was easy to detect defects that broke the continuity of texture.The combination of Smooth-L _(1) loss and GIoU loss improved the regression accuracy on small-size defect samples.Experimental results show that the proposed method performs better than the existing state-of-the-art methods.Moreover,the average detection precision of the proposed method for 11 different classes of surface defects is higher than 98%with few training samples.
作者
张乃雪
钟羽中
赵涛
佃松宜
Zhang Naixue;Zhong Yuzhong;Zhao Tao;Dian Songyi(College of Electrical Engineering,Sichuan University,Chengdu 610065,China)
出处
《计算机应用研究》
CSCD
北大核心
2022年第8期2520-2525,共6页
Application Research of Computers
基金
国家重点研发计划资助项目(2018YFB1307401)。
