钢铁异常检测背景 10篇论文摘要分享 小陈读paper系列

1.Strip steel is an indispensable material in the manufacturing industry and the defects of the surface directly determine the quality. Due to the diversity and complexity of surface defects in intraclass and between interclass, a great deal of manpower and resources have been devoted to surface defect detection.

条纹钢是制造业不可或缺的材料，表面缺陷直接决定了质量。由于类内和类间表面缺陷的多样性和复杂性，大量的人力和资源都致力于表面缺陷检测。

2.Steel surface defect detection is an essential quality control task in manufacturing. As patterns of defects may be viewed as an object, some current defect detection methods, which have achieved promising performance, have been developed based on object-detection models.

钢表面缺陷检测是制造中必不可少的质量控制任务。由于缺陷模式可以被视为对象，一些目前已经基于对象检测模型开发了一些取得了良好的性能的缺陷检测方法。然而，这些缺陷检测方法大多只是加入了额外的重模块来提高准确性。这些方法没有考虑模型的效率或缺陷的特征。

3.Defect classification exhibits great importance in metal surface defect inspection. Most previous defect classification models are based on fully supervised learning, which requires a large amount of training data with image labels. However, collecting defective images in industrial scenarios is quite difficult due to the well-optimized manufacturing techniques

缺陷分类在金属表面缺陷检测中具有重要意义。以前的大多数缺陷分类模型都是基于完全监督的学习，这需要大量带有图像标签的训练数据。然而，由于制造技术的优化良好，在工业场景中收集有缺陷的图像是非常困难的。此外，图像注释也很昂贵且耗时。

4.— As samples of steel defects are industrially limited, it is challenging for most deep learning methods that rely on ample labeled data to identify steel surface defects. Recently, contrastive learning has achieved good performance in natural image classification tasks with few labeled samples, yet two obstacles prevent its effective application to steel surface defect images. One is that due to the presence of inter-class and intra-class similar samples in steel surface defect, the fixed contrast strength in contrastive learning will destroy the potential semantic information of defect samples. Another is that contrastive learning requires a large amount of unlabeled data, whereas steel surface defect samples are insufficient.

由于钢缺陷样品在工业上受到限制，大多数依赖大量标记数据来识别钢表面缺陷的深度学习方法都具有挑战性。近年来，对比学习在标记样本较少的自然图像分类任务中取得了良好的性能，但两个障碍阻碍了其对钢表面缺陷图像的有效应用。一是由于钢表面缺陷中存在类间和类内相似样本，对比学习中的固定对比度强度会破坏缺陷样本的潜在语义信息。另一个是对比学习需要大量的未标记数据，而钢表面缺陷样本不足。

5. The surface defect detection plays an important role in industrial production and directly affects production efficiency and product quality. In this article, we focus on the surface defect detection of heat sink and propose a method based on a Ghost-SE light U-Net (GSLU-Net). The GSLU-Net is a novel combination of the lightweight convolution module, self-attention mechanism, and fully convolutional network (FCN). It has a symmetrical architecture inspired by the U-Net. By introducing the Ghost module, which can generate feature maps with cheap operations, we reduce the computation cost while maintaining high accuracy.

表面缺陷检测在工业生产中发挥着重要作用，直接影响生产效率和产品质量。在本文中，我们专注于散热器的表面缺陷检测，提出了一种基于Ghost-SE光U-Net (GSLU-Net)的方法。GLU-Net 是轻量级卷积模块、自注意力机制和全卷积网络 (FCN) 的新组合。它具有受 U-Net 启发的对称架构。通过引入Ghost模块，该模块可以以廉价的操作生成特征映射，在保持较高的精度的同时，降低了计算成本。

6.Efficiency of surface defect detection is largely improved with the application of convolutional neural networks (CNNs). However, CNNs encounter difficulties in accurately modeling the multiscale features of defects due to the inherent locality of convolution. Additionally, existing methods for cross-level feature fusion introduce cumbersome computational complexity. Consequently, current networks struggle to reconcile location accuracy of multiscale defects with detection efficiency, resulting in poor performance in real-time detection.

随着卷积神经网络 (CNN) 的应用，表面缺陷检测的效率大大提高。然而，由于卷积的固有局部性，cnn在精确建模缺陷多尺度特征方面遇到了困难。此外，现有的跨层特征融合方法引入了繁琐的计算复杂度。因此，目前的网络难以将多尺度缺陷的位置精度与检测效率相协调，导致实时检测性能较差。为此，我们提出了一种用于表面缺陷检测的动态变压器网络，利用基于自我注意的变压器在路由空间中精确捕获远程语义信息。

7.Defect detection in the industry is an essential task in quality inspection. The main target is to classify and localize defects in acquired images. During image acquisition, external noise and diverse background patterns can lead to conflicting information during fusion of network features, which brings certain challenges to detection. In addition, one-stage detectors generate mismatches when performing classification and localization, which can lead to a certain degree of misalignment in prediction. Also, some small defects to be detected in industrial products must be combined with fine-grained information.

工业中的缺陷检测是质量检测中的一项重要任务。主要目标是对获取图像中的缺陷进行分类和定位。在图像采集过程中，外部噪声和不同的背景模式在网络特征融合过程中会导致信息冲突，这给检测带来了一定的挑战。此外，一级检测器在执行分类和定位时产生不匹配，这可能导致预测中的某些错位程度。此外，工业产品中检测到的一些小缺陷必须与细粒度信息相结合。

8.Rain-like layer removal from hot-rolled steel strip surface has been proven to be a workable measure for suppressing the false alarms frequently triggered in automated visual inspection (AVI) instruments. This article extends the scope of the "rain-like layer" from dispersed waterdrops to splashing water streaks and tiny white droplets. And a targeted method with both channel-wise and spatial-wise attention, namely attentive dual residual generative adversarial network (ADRGAN), is proposed.

从热轧钢带表面去除雨状层已被证明是抑制自动目视检查 (AVI) 仪器中经常触发的误报的可行措施。本文扩展了从分散的水滴到飞溅水条纹和微小的白色液滴的“雨样层”的范围。提出了一种同时具有通道和空间注意的有针对性的方法，即注意力双残差生成对抗网络 (ADRGAN)。

9.The worldwide transportation industry relies heav-1ily on shipping containers. Containerization has made it easier2to transfer goods all over the world by guaranteeing cargo3safety while in transit. To ensure the safety of goods during4the transition, shipping containers should be reliable and kept5in healthy conditions. Surface defect inspection of shipping6containers is of great importance to guarantee the quality of7containers. Customs officers must check the surface of shipping8containers as they pass through terminal gates during the9transition. Human visual observation is the basis for the current10inspection method, which is time-consuming, labor-intensive, and11possibly hazardous. The purpose of this research is to present a12deep learning-based framework that can be used in conjunction13with a computer vision technique to successfully and efficiently14inspect corrosion defects on the surface of shipping contain-15ers.

全球运输行业主要依靠航运集装箱。集装箱化通过保证货物在运输过程中的安全，使世界各地的货物转移更容易。为了确保过渡过程中货物的安全，运输集装箱在健康条件下应该是可靠和保存的。航运6集装箱的表面缺陷检测对于保证7个集装箱的质量具有重要意义。海关官员必须检查航运8集装箱的表面，因为它们在9过渡期间通过终端门。人类视觉观察是当前10种检查方法的基础，该方法耗时、劳动密集型和11种可能是危险的。这项研究的目的是提出一个12个基于深度学习的框架，该框架可以与计算机视觉技术相结合，成功地高效地在航运包含-15ers表面的腐蚀缺陷。

10.One of these tasks is to control the surface condition of steel blanks and identify defects. Currently, machine learning methods applied as part of steel slab surface inspection systems require a large number of defect images for training. This in turn increases the time required to collect and markup the training dataset [7,8].

This study considers the possibility of applying synthesized data for semantic segmentation and classification of defects in steel products. The developed approach is supposed to be used in automatic control systems of steel rolling production. These systems include vision-based quality control systems. The task of determining defects on the workpiece surface is complex; it combines several independent vision tasks. First of all, it is necessary to determine the presence of surface defects in the image [ 11]. It is necessary to have a clear idea about the permissible visual deviations, which can lead to false positive recognitions. For example, grease residues, water drops, or fragments of slab markings can be such deviations (Figure 1).

其中一个任务是控制钢毛坯的表面条件并识别缺陷。目前，作为钢板表面检测系统一部分应用的机器学习方法需要大量的缺陷图像进行训练。这反过来又增加了收集和标记训练数据集所需的时间 [7,8]。

本研究考虑了将合成数据应用于钢制品缺陷语义分割和分类的可能性。所开发的方法应用于钢轧制生产的自动控制系统。这些系统包括基于视觉的质量控制系统。确定工件表面缺陷的任务是复杂的;它结合了几个独立的视觉任务。首先，有必要确定图像中表面缺陷的存在[11]。有必要清楚地了解允许的视觉偏差，这可能会导致误报识别。例如，油脂残基、水滴或板状标记的碎片可能是这样的偏差(图1)。