Non-defective Image Research Articles

不変情報クラスタリングと良品画像のみを用いて訓練された欠陥検知用CNNの提案

不変情報クラスタリングと良品画像のみを用いて訓練された欠陥検知用CNNの提案

Vision Transformers for Anomaly Detection and Localisation in Leather Surface Defect Classification Based on Low-Resolution Images and a Small Dataset

Genuine leather manufacturing is a multibillion-dollar industry that processes animal hides from varying types of animals such as sheep, alligator, goat, ostrich, crocodile, and cow. Due to the industry’s immense scale, there may be numerous unavoidable causes of damages, leading to surface defects that occur during both the manufacturing process and the bovine’s own lifespan. Owing to the heterogenous and manifold nature of leather surface characteristics, great difficulties can arise during the visual inspection of raw materials by human inspectors. To mitigate the industry’s challenges in the quality control process, this paper proposes the application of a modern vision transformer (ViT) architecture for the purposes of low-resolution image-based anomaly detection for defect localisation as a means of leather surface defect classification. Utilising the low-resolution defective and non-defective images found in the opensource Leather Defect detection and Classification dataset and higher-resolution MVTec AD anomaly benchmarking dataset, three configurations of the vision transformer and three deep learning (DL) knowledge transfer methods are compared in terms of performance metrics as well as in leather defect classification and anomaly localisation. Experiments show the proposed ViT method outperforms the light-weight state-of-the-art methods in the field in the aspect of classification accuracy. Besides the classification, the low computation load and low requirements for image resolution and size of training samples are also advantages of the proposed method.

Normal-Knowledge-Based Pavement Defect Segmentation Using Relevance-Aware and Cross-Reasoning Mechanisms

Automatic pavement defect segmentation is a big challenge because of the class diversity and extremely random distribution of defects. Most existing approaches focus on supervised strategies to achieve decent performance. Due to the difficulty of getting massive densely annotated samples and the limited prior knowledge of potential defects, these methods have significant bottlenecks in the actual pavement settings. This paper proposes a relevance-aware and cross-reasoning network (RCN) for anomaly segmentation of pavement defects, which can segment defects using merely non-defective images for training. A relevance-aware transformer-based encoder is first devised to model intrinsic interdependencies across local features, thus improving representations of complex non-defective images. Next, a dual decoder strategy is proposed to remap the encoder-generated latent dependencies at the local semantic and global detailed levels, respectively. Specifically, a cross-reasoning refinement module is built in the local decoder to reason the cross-relationship between spatial and channel dimensions. Finally, a context-aware abnormal distillation measurement is developed to evaluate the semantic reconstruction deviations during the inference. Under the guidance of semantic affinity, this measurement allows our model to highlight defective areas adaptively. Extensive experimental results on four datasets indicate that RCN outperforms other leading anomaly segmentation methods.

Anomaly detection from images in pipes using GAN

In recent years, the number of pipes that have exceeded their service life has increased. For this reason, earthworm-type robots equipped with cameras have been developed to perform regularly inspections of sewer pipes. However, inspection methods have not yet been established. This paper proposes a method for anomaly detection from images in pipes using Generative Adversarial Network (GAN). A model that combines f-AnoGAN and Lightweight GAN is used to detect anomalies by taking the difference between input images and generated images. Since the GANs are only trained with non-defective images, they are able to convert an image containing defects into one without them. Subtraction images is used to estimate the location of anomalies. Experiments were conducted using actual images of cast iron pipes to confirm the effectiveness of the proposed method. It was also validated using sewer-ml, a public dataset.

Concatenated Network Fusion Algorithm (CNFA) Based on Deep Learning: Improving the Detection Accuracy of Surface Defects for Ceramic Tile

The low accuracy of detection algorithms is one impediment in detecting ceramic tile’s surface defects online utilizing intelligent detection instead of human inspection. The purpose of this paper is to present a CNFA for resolving the obstacle. Firstly, a negative sample set is generated online by non-defective images of ceramic tiles, and a comparator based on a modified VGG16 extracts a reference image from it. Disguised rectangle boxes, including defective and non-defective, are acquired from the image to be inspected by a detector. A reference rectangle box most similar to the disguised rectangle box is extracted from the reference image. A discriminator is constituted with a modified MobileNetV3 network serving as the backbone and a metric learning loss function strengthening feature recognition, distinguishing the true and false of disguised and reference rectangle boxes. Results exhibit that the discriminator appears to have an accuracy of 98.02%, 13% more than other algorithms. Furthermore, the CNFA performs an average accuracy of 98.19%, and the consumption time of a single image extends by only 64.35 ms, which has little influence on production efficiency. It provides a theoretical and practical reference for surface defect detection of products with complex and changeable textures in industrial environments.

Surface Defect Detection for Mobile Phone Back Glass Based on Symmetric Convolutional Neural Network Deep Learning

Defect detection based on machine vision and machine learning techniques has drawn much attention in recent years. Deep learning is very suitable for such segmentation and detection tasks and has become a promising research area. Surface quality inspection is essentially important in the manufacturing of mobile phone back glass (MPBG). Different types of defects are produced because of the imperfection of the manufacturing technique. Unlike general transparent glass, screen printing glass has totally different reflection and scattering characteristics, which means the traditional dark-field imaging system is not suitable for this task. Meanwhile, the imaging system requires high resolution since the minimum defect size can be 0.005 mm2. According to the imaging characteristics of screen printing glass, this paper proposes a coaxial bright-field (CBF) imaging system and low-angle bright-field (LABF) imaging system, and 8K line-scan complementary metal oxide semiconductor(CMOS) cameras are utilized to capture images with the resolution size of 16,000*8092. The CBF system is applied for the weak-scratch and discoloration defects while the LABF system is applied for the dent defect. A symmetric convolutional neural network composed of encoder and decoder structures is proposed based on U-net, which produces a semantic segmentation with the same size as the original input image. More than 10,000 original images were captured, and more than 30,000 defective and non-defective images were manually annotated in the glass surface defect dataset (GSDD). Verified by the experiments, the results showed that the average precision reaches more than 91% and the average recall rate reaches more than 95%. The method is very suitable for the surface defect inspection of screen printing mobile phone back glass.

Convolutional Neural Network Based Surface Inspection System for Non-patterned Welding Defects

In this paper, we propose a convolutional neural network (CNN) based method that inspects non-patterned welding defects (craters, pores, foreign substances and fissures) on the surface of the engine transmission using a single RGB camera. The proposed method consists of two steps: first, extracting the welding area to be inspected from the captured image, and then determining whether the extracted area includes defects. In the first step, to extract the welding area from the captured image, a CNN based approach is proposed to detect a center of the engine transmission in the image. In the second stage, the extracted area is identified by another CNN as defective or non-defective. To train the second stage CNN stably, we propose a class-specific batch sampling method. With our sampling method, biased learning caused by data imbalance (number of collected defective images is much less than that of non-defective images) is effectively prevented. We evaluated our system with a large amount of samples (about 32,000 images) collected manually from the production line, and our system shows a remarkable performance in all experiments.

A real-time computer vision-based platform for fabric inspection part 1: algorithm

Developing an efficient real-time detection algorithm is quite important for an automated inspection system. This paper presents a practical method based on local singular value decomposition (SVD) and normalised cross-correlation (NCC) for real-time defect detection in woven fabrics. As fabric-textured images exhibit high periodicity among the repeated sub-patterns, non-defective or normal image samples (image patches) can be efficiently approximated as a linear combination of the basis vectors (BVs) obtained via SVD. Since these BVs are recovered from normal samples, they will only capture the key structural features of the non-defective images. When using the BVs to model new samples, we can expect defective or abnormal samples with structural features not found in normal cases will incur substantial approximation errors. Therefore, complex defect detection can be converted to a template matching problem, where the robust NCC is utilised to measure disparities between the original and its approximation for defect identification. Experimental results on various real-world fabrics exhibit accurate defect detection with low false alarm rate, and we also conduct a comparison with a feature extraction-based method to further confirm the effectiveness of our algorithm.

Ｘ線式はんだ付検査装置の開発

This paper describes an automatic solder joint inspection system by X-ray imaging. The following were developed for highly reliable inspection of J-lead and gull-wing lead solder joints : (1) An automatic inspection method for double side-mounted PCBs (Printed Circuit Boards) by micro-focus X-ray imaging and board tilting, (2) Enhancement of X-ray image of solder joints by logarithmic transformation of a detected image, (3) Fillet judgement by comparison between a detected image and a number of typical nondefective images selected by a clustering procedure, (4) Bridge determination by multi-step thresholding and comparison between pattern frequency distribution of a detected image and that of a nondefective image. The inspection system developed with these technologies had a defect detection rate of 100% with the smallest detectable defect being 50 μm.